Mastering Unix Shell Scripting
Randal K. Michael

Free & Share & Open

The
Dear Valued Customer,

WILEY advantage

We realize you’re a busy professional with deadlines to hit. Whether your goal is to learn a new technology or solve a critical problem, we want to be there to lend you a hand. Our primary objective is to provide you with the insight and knowledge you need to stay atop the highly competitive and everchanging technology industry. Wiley Publishing, Inc., offers books on a wide variety of technical categories, including security, data warehousing, software development tools, and networking — everything you need to reach your peak. Regardless of your level of expertise, the Wiley family of books has you covered. • For Dummies – The fun and easy way to learn • The Weekend Crash Course –The fastest way to learn a new tool or technology • Visual – For those who prefer to learn a new topic visually • The Bible – The 100% comprehensive tutorial and reference • The Wiley Professional list – Practical and reliable resources for IT professionals The book you hold now, Mastering Unix Shell Scripting, is the first book to provide end-to-end scripting solutions that will solve real-world system administration problems for those who have to automate these often complex and repetitive tasks. Starting with a sample task and targeting the most common Unix systems: Solaris, Linux, AIX, and HP-UX with specific command structures, this book will save precious time with hands-on detail. The companion Web site contains all the timesaving scripts from the book. Our commitment to you does not end at the last page of this book. We’d want to open a dialog with you to see what other solutions we can provide. Please be sure to visit us at www.wiley.com/compbooks to review our complete title list and explore the other resources we offer. If you have a comment, suggestion, or any other inquiry, please locate the “contact us” link at www.wiley.com. Thank you for your support and we look forward to hearing from you and serving your needs again in the future. Sincerely,

Richard K. Swadley Vice President & Executive Group Publisher Wiley Technology Publishing

Free & Share & Open

Mastering Unix Shell Scripting

Free & Share & Open

Mastering Unix Shell Scripting
Randal K. Michael

Free & Share & Open

Publisher: Robert Ipsen Executive Editor: Carol Long Developmental Editor: Scott Amerman Managing Editor: Angela Smith Text Design & Composition: Wiley Composition Services This book is printed on acid-free paper. ∞ Copyright © 2003 by Randal K. Michael. All rights reserved. Published by Wiley Publishing, Inc., Indianapolis, Indiana Published simultaneously in Canada No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470. Requests to the Publisher for permission should be addressed to the Legal Department, Wiley Publishing, Inc., 10475 Crosspoint Blvd., Indianapolis, IN 46256, (317) 572-3447, fax (317) 572-4447, E-mail: PERMCOORDINATOR@WILEY.COM. Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages. For general information on our other products and services please contact our Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002. Trademarks: Wiley, the Wiley Publishing logo and related trade dress are trademarks or registered trademarks of Wiley Publishing, Inc., in the United States and other countries, and may not be used without written permission. All other trademarks are the property of their respective owners. Wiley Publishing, Inc., is not associated with any product or vendor mentioned in this book. Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books. Library of Congress Cataloging-in-Publication Data: ISBN: 0-471-21821-9 Printed in the United States of America 10 9 8 7 6 5 4 3 2 1

This book is dedicated to My Wife Robin, and the girls, Andrea and Ana

Free & Share & Open

Acknowledgments

The information that I gathered together in this book is the result of working with some of the most talented UNIX professionals on the topic. I have enjoyed every minute of my association with these UNIX gurus and it has been my pleasure to have the opportunity to gain so much knowledge from the pros. I want to thank every one of these experts for asking and answering questions over the last fifteen years. If my brother, Jim, had not kept telling me, “you should write a book,” after querying me for UNIX details on almost a weekly basis, I doubt this book would have ever been written. So, thanks Jim! I especially want to thank Jack Renfro at Daimler/Chrysler Corporation for giving me my first shell scripting project so long ago. I had to start with the man pages, but that is how I learned to dig deep to get an answer. Since then I have been on a mission to automate, through shell scripting, everything on every system that I come in contact with. I certainly value the years that I was able to work with Jack. I must also thank the talented people at Wiley Publishing. Margaret Eldridge started me on this project by letting me do my own thing, and Carol Long kept me going. Scott Amerman kept me on schedule, and Angela Smith did the edits that make my writing flow with ease. It has been a valuable experience for me to work with such a fine group of professionals at Wiley. I also want to thank Carole McClendon at Waterside Productions for all of the support on this project. Carole is the best Agent that anyone could ever ask for. She is a true professional with the highest ethics. Of course my family had a lot to do with my success on this and every project. I want to thank Mom, Gene, Jim, Marcia, Rusty, Mallory, and Anica. I want to thank my Wife Robin for her understanding and support. The girls, Andrea and Ana, always keep a smile on my face, and Steve is always on my mind. I could not have written this book without the support of all of these people and the many others that remain unnamed. It has been an honor!

vii

Free & Share & Open

Contents

Acknowledgments Introduction Chapter 1 Scripting Quick Start and Review Case Sensitivity Unix Special Characters Shells Shell Scripts Functions A Function Has the Form Running a Shell Script Declare the Shell in the Shell Script Comments and Style in Shell Scripts Control Structures Using break, continue, exit, and return Here Document Syntax for a Here Document Shell Script Commands Symbol Commands Variables Command-Line Arguments Shift Command Special Parameters $* and $@ Special Parameter Definitions Double Quotes “, Forward Tics ‘, and Back Tics ` Math in a Shell Script Operators Built-in Mathematical Functions

vii xix 1 1 2 2 2 2 3 3 3 4 6 9 9 9 10 13 13 13 14 15 15 16 17 17 18

ix

Free & Share & Open

x

Contents File Permissions, suid and sgid Programs chmod Command Syntax for Each Purpose Running Commands on a Remote Host Setting Traps User Information Commands who Command w Command last Command ps Command Communicating with Users Uppercase or Lowercase Text for Easy Testing Check the Return Code Time-Based Script Execution cron tables Cron Table Entry Syntax Wildcards at Command Output Control Silent Running Using getopts to Parse Command-Line Arguments Making a Co-Process with Background Function Catching a Delayed Command Output Fastest Ways to Process a File Line -by Line Mail Notification Techniques Using the mail and mailx Commands Using the sendmail Command to Send Outbound Mail Creating a Progress Indicator A Series of Dots A Rotating Line Creating a Psuedo-Random Number Checking for Stale Disk Partitions in AIX Automated Host Pinging Highlighting Specific Text in a File Keeping the Printers Printing AIX “Classic” Printer Subsystem System V Printing Automated FTP File Transfer Capturing a List of Files Larger than $MEG Capturing a User’s Keystrokes Using the bc Utility for Floating-Point Math Number Base Conversions Using the typeset Command Using the printf Command Create a Menu with the select Command Sending Pop-Up Messages to Windows Removing Repeated Lines in a File Removing Blank Lines from a File 18 19 20 21 22 22 22 22 23 23 24 25 27 27 27 28 28 28 28 29 30 32 33 34 34 34 35 35 35 36 37 37 38 38 38 39 39 39 40 40 41 41 41 42 43 43 44

Contents Testing for a Null Variable Directly Access the Value of the Last Positional Parameter, $# Remove the Columns Heading in a Command Output Arrays Loading an Array Testing a String Summary Chapter 2 Twelve Ways to Process a File Line by Line Command Syntax Using File Descriptors Creating a Large File to Use in the Timing Test Twelve Methods to Parse a File Line by Line Method 1: cat $FILENAME | while read LINE Method 2: while read $FILENAME from Bottom Method 3: while_line_LINE_Bottom Method 4: cat $FILENAME | while LINE=`line` Method 5: cat $FILENAME | while line LINE Method 6: while LINE=`line` from the Bottom Method 7: cat $FILENAME | while LINE=$(line) Method 8: while LINE=$(line) from the Bottom Method 9: while read LINE Using File Descriptors Method 10: while LINE=’line’ Using File Descriptors Method 11: while LINE=$(line) Using File Descriptors Method 12: while line LINE Using File Descriptors Timing Each Method Timing Script Timing Data for Each Method Timing Command Substitution Methods Summary Automated Event Notification Basics of Automating Event Notification Using the mail and mailx Commands Problems with Outbound Mail Create a “Bounce” Account with a .forward File Using the sendmail Command to Send Outbound Mail Dial-Out Modem Software SNMP Traps Summary Progress Indicator Using a Series of Dots, a Rotating Line, or a Countdown to Zero Indicating Progress with a Series of Dots Indicating Progress with a Rotating Line Creating a Countdown Indicator Other Options to Consider Summary 44 45 45 46 46 47 51 53 53 54 54 56 57 58 58 59 60 61 61 62 63 64 65 66 66 67 73 77 78 79 79 80 82 82 83 84 85 86 87 87 89 91 95 96

xi

Chapter 3

Chapter 4

Free & Share & Open

xii

Contents Chapter 5 File System Monitoring Syntax Adding Exceptions Capability to Monitoring The Exceptions File Using the MB of Free Space Method Using MB of Free Space with Exceptions Percentage Used—MB Free and Large Filesystems Running on AIX, Linux, HP-UX, and Solaris Command Syntax and Output Varies between Operating Systems Other Options to Consider Event Notification Automated Execution Modify the egrep Statement Summary Monitoring Paging and Swap Space Syntax AIX lsps Command HP-UX swapinfo Command Linux free Command Solaris swap Command Creating the Shell Scripts AIX Paging Monitor HP-UX Swap Space Monitor Linux Swap Space Monitor Solaris Swap Space Monitor All-in-One Paging and Swap Space Monitor Other Options to Consider Event Notification Log File Scheduled Monitoring Summary Monitoring System Load Syntax Syntax for uptime AIX HP-UX Linux Solaris What Is the Common Denominator? Scripting an uptime Field Test Solution Syntax for iostat AIX HP-UX 97 98 103 103 110 113 118 128
130 143 143 143 144 144

Chapter 6

145 146 146 147 148 148 149 149 155 160 164 169 176 177 177 177 177 179 180 180 180 181 182 183 183 184 186 186 186

Chapter 7

Contents
Linux Solaris What Is the Common Denominator? Syntax for sar AIX HP-UX Linux Solaris What Is the Common Denominator? Syntax for vmstat AIX HP-UX Linux Solaris What Is the Common Denominator? 187 187 187 188 188 189 189 190 190 191 191 191 192 192 192

xiii

Scripting the Solutions Using uptime to Measure the System Load Scripting with the uptime Command Using sar to Measure the System Load Scripting with the sar Command Using iostat to Measure the System Load Scripting with the iostat Command Using vmstat to Measure the System Load Scripting with the vmstat Command Other Options to Consider Stop Chasing the Floating uptime Field Try to Detect Any Possible Problems for the User Show the User the Top CPU Hogs Gathering a Large Amount of Data for Plotting Summary Chapter 8 Process Monitoring and Enabling Preprocess, Startup, and Postprocess Events Syntax Monitoring for a Process to Start Monitoring for a Process to End Monitor and Log as a Process Starts and Stops Timed Execution for Process Monitoring, Showing each PID, and Time Stamp with Event and Timing Capability Other Options to Consider Common Uses Modifications to Consider Summary Monitoring Processes and Applications Monitoring Local Processes Remote Monitoring with Secure Shell Checking for Active Oracle Databases Checking If the HTTP Server/Application Is Working

193 194 194 197 198 203 203 208 208 212 212 213 213 214 214 215 216 216 218 223 228 248 248 248 249 251 252 254 256 259

Chapter 9

Free & Share & Open

xiv

Contents Other Things to Consider Application APIs and SNMP Traps Summary Chapter 10 Creating Pseudo-Random Passwords Randomness Creating Pseudo-Random Passwords Syntax Arrays Loading an Array Building the Password Creation Script Order of Appearance Define Variables Define Functions Testing and Parsing Command-Line Arguments Beginning of Main Setting a Trap Checking for the Keyboard File Loading the “KEYS” Array Using the LENGTH Variable to Build a Loop List Building a New Pseudo-Random Password Printing the Manager’s Password Report for Safe Keeping Other Options to Consider Password Reports? Which Password? Other Uses? Summary Chapter 11 Monitor for Stale Disk Partitions AIX Logical Volume Manager (LVM) The Commands and Methods Disk Subsystem Commands Method 1: Monitoring for Stale PPs at the LV Level Method 2: Monitoring for Stale PPs at the PV Level Method 3: VG, LV, and PV Monitoring with a resync Other Options to Consider SSA Disks Log Files Automated Execution Event Notification Summary Chapter 12 Automated Hosts Pinging with Notification Syntax Creating the Shell Script Define the Variables Creating a Trap The Whole Shell Script 260 261 261 263 263 264 264 265 265 266 266 266 267 275 279 280 280 280 281 282 283 294 294 295 295 295 297 298 298 298 299 304 308 315 315 316 316 316 317 319 320 321 321 323 324

Contents Other Options to Consider $PINGLIST Variable Length Limit Problem Ping the /etc/hosts File Instead of a List File Logging Notification of “Unknown host” Notification Method Automated Execution Using a Cron Table Entry Summary Chapter 13 Taking a System Snapshot Syntax Creating the Shell Script Other Options to Consider Summary Chapter 14 Compiling, Installing, Configuring, and Using sudo The Need for sudo Downloading and Compiling sudo Compiling sudo Configuring sudo Using sudo Using sudo in a Shell Script The sudo Log File Summary Chapter 15 hgrep: Highlighted grep Script Reverse Video Control Building the hgrep.ksh Shell Script Other Options to Consider Other Options for the tput Command Summary Chapter 16 Print Queue Hell: Keeping the Printers Printing System V versus BSD Printer Subsystems AIX Print Control Commands Classic AIX Printer Subsystem System V Printing on AIX More System V Printer Commands HP-UX Print Control Commands Linux Print Control Commands Controlling Queuing and Printing Individually Solaris Print Control Commands More System V Printer Commands Putting It All Together Other Options to Consider Logging Exceptions Capability Maintenance Scheduling Summary 332 332 333 333 334 334 335 335 337 338 340 367 367 369 369 370 371 378 384 385 389 390 391 392 393 400 400 401 403 404 404 404 408 412 414 417 422 425 429 431 438 439 439 439 439 439

xv

Free & Share & Open

xvi

Contents Chapter 17 Automated FTP Stuff Syntax Automating File Transfers and Remote Directory Listings Using FTP for Directory Listings on a Remote Machine Getting One or More Files from a Remote System Pre and Post Events Script in Action Putting One or More Files to a Remote System Replacing Hard-Coded Passwords with Variables Example of Detecting Variables in a Script’s Environment Modifying Our FTP Scripts to Use Password Variables Other Options to Consider Use Command-Line Switches to Control Execution Keep a Log of Activity Add a Debug Mode to the Scripts Summary Chapter 18 Finding “Large” Files Syntax Creating the Script Other Options to Consider Summary Chapter 19 Monitoring and Auditing User Key Strokes Syntax Scripting the Solution Logging User Activity Starting the Monitoring Session Where Is the Repository? The Scripts Logging root Activity Some sudo Stuff Monitoring Other Administration Users Other Options to Consider Emailing the Audit Logs Compression Need Better Security? Inform the Users Sudoers File Summary Chapter 20 Turning On/Off SSA Identification Lights Syntax Translating an hdisk to a pdisk Identifying an SSA Disk The Scripting Process Usage and User Feedback Functions Control Functions The Full Shell Script 441 441 444 444 446 449 449 450 452 453 456 463 463 463 463 464 465 466 466 472 473 475 476 477 478 479 479 480 483 486 489 492 493 493 493 493 494 494 495 496 496 496 497 497 501 507

Contents Other Things to Consider Error Log Cross-Reference Root Access and sudo Summary Chapter 21 Pseudo-Random Number Generation What Makes a Random Number? The Methods Method 1: Creating Numbers between 0 and 32,767 Method 2: Creating Numbers between 1 and a User-Defined Maximum Method 3: Fixed-Length Numbers between 1 and a User-Defined Maximum Why Pad the Number with Zeros the Hard Way? Shell Script to Create Pseudo-Random Numbers Creating Unique Filenames Summary Chapter 22 Floating-Point Math and the bc Utility Syntax Creating Some Shell Scripts Using bc Creating the float_add.ksh Shell Script Testing for Integers and Floating-Point Numbers Building a Math Statement for the bc Command Using a Here Document Creating the float_subtract.ksh Shell Script Using getopts to Parse the Command Line Building a Math Statement String for bc Here Document and Presenting the Result Creating the float_multiply.ksh Shell Script Parsing the Command Line for Valid Numbers Creating the float_divide.ksh Shell Script Creating the float_average.ksh Shell Script Other Options to Consider Remove the Scale from Some of the Shell Scripts Create More Functions Summary Chapter 23 Scripts for Number Base Conversions Syntax Example 23.1: Converting from Base 10 to Base 16 Example 23.2: Converting from Base 8 to Base 16 Example 23.3 Converting Base 10 to Octal Example 23.4 Converting Base 10 to Hexadecimal Scripting the Solution Base 2 (binary) to Base 16 (hexadecimal) Shell Script Base 10 (Decimal) to Base 16 (Hexadecimal) Shell Script 520 520 520 520 521 523 523 524 525
526 527 529

xvii

530 535 543 545 545 546 546 552 554 555 556 561 563 564 565 570 573 580 582 582 582 583 585 585 586 586 587 587 587 587 590

Free & Share & Open

xviii Contents
Script to Create a Software Key Based on the Hexadecimal Representation of an IP Address Script to Translate between Any Number Base Using getopts to Parse the Command Line Example 23.5 Correct Usage of the Equate_any_base.ksh Shell Script Example 23.6 Incorrect Usage of the Equate_any_base.ksh Shell Script Continuing with the Script Beginning of Main

594 597 602 603 603 604 606

Other Options to Consider Software Key Shell Script Summary Chapter 24 Menu Program Suitable for Operations Staff Reverse Video Syntax Creating the Menu Creating a Message Bar for Feedback From the Top Other Options to Consider Shelling Out to the Command Line Good Candidate for Using sudo Summary Chapter 25 Sending Pop-Up Messages from Unix to Windows About Samba and the smbclient Command Syntax Building the broadcast.ksh Shell Script Sending a Message to All Users Adding Groups to the Basic Code Adding the Ability to Specify Destinations Individually Using getopts to Parse the Command Line Testing User Input Testing and Prompting for WINLIST Data Testing and Prompting for Message Data Sending the Message Putting It All Together Watching the broadcast.ksh Script in Action Downloading and Installing Samba Testing the smbclient Program the First Time Other Options to Consider Producing Error Notifications Add Logging of Unreachable Machines Create Two-Way Messanging Summary Appendix A What’s on the Web Site Index

608 608 608 609 610 610 611 616 617 618 618 618 619 619 620 621 621 623 623 624 627 627 628 629 630 640 642 643 644 645 645 645 645 647 663

Introduction

In Unix there are many ways to accomplish a given task. Given a problem to solve, we may be able to get to a solution in any number of ways. Of course, some will be more efficient, be more readable, use less disk space or memory, may or may not give the user feedback on what is going on or give more accurate details and more precision to the result. In this book we are going to step through every detail of writing a shell script to solve real-world Unix problems and tasks. The shell scripts range from using a pseudorandom number generator to create pseudo-random passwords to checking for full filesystems on Unix machines and to sending pop-up messages to Windows desktops. The details required to write these shell scripts include using good style and providing good comments throughout the shell script by describing each step. Other details include combining many commands into just one command statement when desirable, separating commands on several lines when readability and understanding of the concept may be diminished, and making a script readable and easy to maintain. We will see the benefit of using variables and files to store data, show methods to strip out unwanted or unneeded data from a command output, and format the data for a particular use. Additionally, we are going to show how to write and include functions in our shell scripts and demonstrate the benefits of functions over a shell script written without functions. This book is intended for any flavor of Unix, but its emphasis includes AIX, Linux, HP-UX, and Solaris operating systems. Most every script in the book is also included on the book’s companion Web site (www.wiley.com/compbooks/michael). Many of the shell scripts are rewritten for each different operating system, when it is necessary. Other shell scripts are not platform dependent. These script rewrites are sometimes needed because command syntax and output vary, sometimes in a major way, between Unix flavors. The variations are sometimes as small as pulling the data out of a different column or using a different command switch, or they can be as major as putting several commands together to accomplish the same task to get similar output or result on different flavors of Unix. In each chapter we start with the very basic concepts and work our way up to some very complex and difficult concepts. The primary purpose of a shell script is automating repetitive and complex functions. This alleviates keystroke errors and allows for timescheduled execution of the shell script. It is always better to have the system tell us that

xix

Free & Share & Open

xx

Introduction it has a problem than to find out too late to be proactive. This book will help us to be more proactive in our dealings with the system. At every level we will gain more knowledge to allow us to move on to ever increasingly complex ideas with ease. We are going to show different ways to solve our real-world example tasks. There is not just one correct way to solve a challenge, and we are going to look at the pros and cons of attacking a problem in various ways. Our goal is to be confident and flexible problem solvers. Given a task, we can solve it in any number of ways, and the solution will be intuitively obvious when you complete this book.

Overview of the Book and Technology
This book is intended as a learning tool and study guide to learn how to write shell scripts to solve a multitude of problems by starting with a clear goal. While studying with this book we will cover most shell scripting techniques about seven times, each time from a different angle, solving a different problem. I have found this learning technique to work extremely well for retention of the material to memory. I urge everyone to read this book from cover to cover to get the maximum benefit. Every script is written using Korn shell, which is the industry standard for scripting solutions in Unix, although some may argue this point. There are several versions of the Korn shell shipped with Unix, depending on the Unix operating system (OS) and the version of the OS release. I have found that the shell scripts in this book will run on any of the Korn shell versions without any modification. This book goes from some trivial task solutions to some rather advanced concepts that Systems Administrators will benefit from, and a lot of stuff in between. There are several chapters for each level of complexity scattered throughout this book. The shell scripts presented in this book are complete shell scripts, which is one of the things that sets this book apart from other shell scripting books on the market. The solutions are explained thoroughly, with each part of the shell script explained in minute detail down to the philosophy and mindset of the author.

How This Book Is Organized
Each chapter starts with a typical Unix challenge that occurs every day in the computing world. With each challenge we define a specific goal and start the shell script by defining the correct command syntax to solve the problem. When we have a goal and the command syntax, we start building the shell script around the commands. The next step is to filter the command(s) output to strip out the unneeded data, or we may decide to just extract the data we need from the output. If the syntax varies between Unix flavors we show the correct syntax to get the same, or a similar, result. When we get to this point we go further to build options into the shell script to give the end user more flexibility on the command line. When a shell script has to be rewritten for each operating system, a combined shell script is shown at the end of the chapter that joins the Unix flavor differences together into one shell script that will run on all of the OS flavors. To do this last step we query the system for the Unix flavor using the uname command. By knowing the flavor of the operating system we are able to execute the proper commands for each Unix flavor

Introduction by using a simple case statement. If this is new to you, do not worry; everything is explained throughout the book in detail. Each chapter targets a different real-world problem. Some challenges are very complex, while others are just interesting to play around with. Some chapters hit the problem from several different angles in a single chapter, and others leave you the challenge to solve on your own—of course, with a few hints to get you started. Each chapter solves the challenge presented and can be read as a single unit without referencing other chapters in the book. Some of the material, though, is explained in great detail in one chapter and lightly covered in other chapters. Because of this variation we recommend that you start at the beginning of the book and read and study every chapter to the end of the book because this is a learning experience!

xxi

Who Should Read This Book
This book is intended for anyone who works with Unix on a daily basis from the command line. The topics studied in the book are mainly for Unix professionals— Programmers, Programmer-Analysts, System Operators, Systems Administrators, and anyone who is interested in getting ahead in the support arena. Beginners will get a lot out of this book, too, but some of the material may be a little high level, so a basic Unix book may be needed to answer some questions. Everyone should have a good working knowledge of common Unix commands before starting this book, because we do not explain common Unix commands at all. I started my career in Unix by learning on the job how to be a Systems Operator. I wish I had a book like this when I started. Having this history I wanted others to get a jump start on their careers. I wrote this book with the knowledge that I was in your shoes at one time, and I remember that I had to learn everything from the man pages, one command at a time. Use this book as a study guide, and you will have a jump start to get ahead quickly in the Unix world, which is getting bigger all of the time.

Tools You Will Need
To get the most benefit from this book you need access to a Unix machine, preferably with AIX, HP-UX, Linux, or Solaris installed. You can run Linux and Solaris on standard PC hardware, and it is relatively inexpensive. It is a good idea to make your default shell environment the Korn shell (ksh); the standard shell on Linux is the Bourne Again shell (bash) shell, and some others use Bourne shell (sh) as the default. You can find your default shell by entering echo $SHELL from the command line. None of the shell scripts in this book requires a graphical terminal, but it sure does not hurt to have GNOME, CDE, KDE2, or X-Windows running. This way you can work in multiple windows at the same time and cut and paste code between windows. You also need a text editor that you are comfortable using. Most Unix operating systems come with the vi editor, and a lot also include emacs. Remember that the editor must be a text editor that stores files in a standard ANSII format. The CDE and other X-editors work just fine, too. You will also need some time, patience, and an open, creative mind that is ready to learn.

Free & Share & Open

xxii

Introduction
Another thing to note is that all of the variables used in the shell scripts and functions in this book are in uppercase. I did this because it is much easier to follow along with a shell script if you know quickly where the variables are located in the code. When you write your own shell scripts, please use lowercase for all shell script and function variables. The reason this is important is that the operating system, and applications, use environment variables that are in uppercase. If you are not careful, you can overwrite a critical system or application variable with your own value and hose up the system; however this is dependent on the scope of where the variable is visible in the code. Just a word of warning, be careful with uppercase variables!

What’s on the Web Site
On the book’s companion Web site, www.wiley.com/compbooks/michael, all of the shell scripts and most of the functions that are studied in the book can be found. The functions are easy to cut and paste directly into your own shell scripts to make the scripting process a little easier. Additionally, there is a shell script stub that you can copy to another filename. This script stub has everything to get you started writing quickly. The only thing you need to do is fill in the fields for the following: Script Name, Author, Date, Version, Platform, Purpose, and Rev List, when revisions are made. There is a place to define variables and functions, and then you have a “BEGINNNG OF MAIN” section to start the main body of the shell script.

Summary
This book is for learning how to be creative, proactive, and a professional problem solver. Given a task, the solution will be intuitively obvious to you on completion of this book. This book will help you attack problems logically and present you with a technique of building on what you know. With each challenge presented you will see how to take the basic syntax and turn it into the basis for a shell scripting solution. We always start with the basics and build more and more logic into the solution before we add other options the end user can use for more flexibility. Speaking of end users, we must always keep our users informed about how processing is progressing. Giving a user a blank screen to look at is the worst thing that you can do, so for this we can create progress indicators. You will learn how to be proactive by building tools that monitor for specific situations that indicate the beginning stages of an upcoming problem. This is where knowing how to query the system puts you ahead of the rest of your staff. With the techniques presented in this book, you will learn. You will learn about problem resolution. You will learn about starting with what you know about a situation and building a solution effectively. You will learn how to make a single shell script work on other platforms without further modification. You will learn how to be proactive. You will learn how to write a shell script that is easily maintained. You will learn how to use plenty of comments in a shell script. You will learn how to write a shell script that is easy to read and follow through the logic. Basically, you will learn to be an effective problem solver where the solution to any challenge is intuitively obvious!

CHAPTER

1
Scripting Quick Start and Review

We are going to start out by giving a very targeted refresher course. The topics that follow are short explanations of techniques that we always have to search the book to find; here they are all together in one place. The explanations range from showing the fastest way to process a file line by line to the simple matter of case sensitivity of Unix and shell scripts. This should not be considered a full and complete list of scripting topics, but it is a very good starting point and it does point out a sample of the topics covered in the book. For each topic listed in this chapter there is a very detailed explanation later in the book. I urge everyone to study this entire book. Every chapter hits a different topic using a different approach. The book is written this way to emphasize that there is never only one technique to solve a challenge in Unix. All of the shell scripts in this book are realworld examples of how to solve a problem. Thumb through the chapters, and you can see that I tried to hit most of the common (and some uncommon!) tasks in Unix. All of the shell scripts have a good explanation of the thinking process, and we always start out with the correct command syntax for the shell script targeting a specific goal. I hope you enjoy this book as much as I enjoyed writing it. Let’s get started!

Case Sensitivity
Unix is case sensitive. Because Unix is case sensitive our shell scripts are also case sensitive.

1

Free & Share & Open

2

Chapter 1

Unix Special Characters
All of the following characters have a special meaning or function. If they are used in a way that their special meaning is not needed then they must be escaped. To escape, or remove its special function, the character must be immediately preceded with a backslash, \, or enclosed within ‘ ‘ forward tic marks (single quotes).
\ ( ; # $ ? & * ( ) [ ] ` ‘ “ +

Shells
A shell is an environment in which we can run our commands, programs, and shell scripts. There are different flavors of shells, just as there are different flavors of operating systems. Each flavor of shell has its own set of recognized commands and functions. This book works entirely with the Korn shell.
Korn Shell /bin/ksh OR /usr/bin/ksh

Shell Scripts
The basic concept of a shell script is a list of commands, which are listed in the order of execution. A good shell script will have comments, preceded by a pound sign, #, describing the steps. There are conditional tests, such as value A is greater than value B, loops allowing us to go through massive amounts of data, files to read and store data, and variables to read and store data, and the script may include functions. We are going to write a lot of scripts in the next several hundred pages, and we should always start with a clear goal in mind. By clear goal, we have a specific purpose for this script, and we have a set of expected results. We will also hit on some tips, tricks, and, of course, the gotchas in solving a challenge one way as opposed to another to get the same result. All techniques are not created equal. Shell scripts and functions are both interpreted. This means they are not compiled. Both shell scripts and functions are ASCII text that is read by the Korn shell command interpreter. When we execute a shell script, or function, a command interpreter goes through the ASCII text line by line, loop by loop, test by test and executes each statement, as each line is reached from the top to the bottom.

Functions
A function is written in much the same way as a shell script but is different in that it is defined, or written, within a shell script, most of the time, and is called within the script. This way we can write a piece of code, which is used over and over, just once and use it without having to rewrite the code every time. We just call the function instead.

Scripting Quick Start and Review
We can also define functions at the system level that is always available in our environment, but this is a later topic for discussion.

3

A Function Has the Form function function_name { commands to execute }

or function_name () { commands to execute }

When we write functions into our scripts we must remember to declare, or write, the function before we use it: The function must appear above the command statement calling the function. We can’t use something that does not yet exist.

Running a Shell Script
A shell script can be executed in the following ways: ksh shell_script_name

will create a Korn shell and execute the shell_script_name in the newly created Korn shell environment. shell_script_name will execute shell_script_name if the execution bit is set on the file (see the man page on the chmod command). The script will execute in the shell that is declared on the first line of the shell script. If no shell is declared on the first line of the shell script, it will execute in the default shell, which is the user’s system-defined shell. Executing in an unintended shell may result in a failure and give unpredictable results.

Declare the Shell in the Shell Script
Declare the shell! If we want to have complete control over how a shell script is going to run and in which shell it is to execute, we MUST declare the shell in the very first line

Free & Share & Open

4

Chapter 1 of the script. If no shell is declared, the script will execute in the default shell, defined by the system for the user executing the shell script. If the script was written, for example, to execute in Korn shell ksh, and the default shell for the user executing the shell script is the C shell csh, then the script will most likely have a failure during execution. To declare a shell, one of the declaration statements in Table 1.1 must appear on the very first line of the shell script:

N OT E This book uses only the Korn shell, #!/usr/bin/ksh OR #!/bin/ksh.

Comments and Style in Shell Scripts
Making good comments in our scripts is stressed throughout this book. What is intuitively obvious to us may be total Greek to others who follow in our footsteps. We have to write code that is readable and has an easy flow. This involves writing a script that is easy to read and easily maintained, which means that it must have plenty of comments describing the steps. For the most part, the person who writes the shell script is not the one who has to maintain it. There is nothing worse than having to hack through someone else’s code that has no comments to find out what each step is supposed to do. It can be tough enough to modify the script in the first place, but having to figure out the mind set of the author of the script will sometimes make us think about rewriting the entire shell script from scratch. We can avoid this by writing a clearly readable script and inserting plenty of comments describing what our philosophy is and how we are using the input, output, variables, and files. For good style in our command statements, we need it to be readable. For this reason it is sometimes better, for instance, to separate a command statement onto three separate lines instead of stringing, or piping, everything together on the same line of code; in some cases, it is more desirable to create a long pipe. In some cases, it may be just too difficult to follow the pipe and understand what the expected result should be for a new script writer. And, again, it should have comments describing our thinking step by step. This way someone later will look at our code and say, “Hey, now that’s a groovy way to do that.”

Table 1.1

Different Types of Shells to Declare OR OR OR OR #!/bin/sh #!/bin/ksh #!/bin/csh #!/bin/bash Declares a Bourne shell Declares a Korn shell Declares a C shell Declares a Bourne-Again shell

#!/usr/bin/sh #!/usr/bin/ksh #!/usr/bin/csh #!/usr/bin/bash

Scripting Quick Start and Review
Command readability and step-by-step comments are just the very basics of a wellwritten script. Using a lot of comments will make our life much easier when we have to come back to the code after not looking at it for six months, and believe me, we will look at the code again. Comment everything! This includes, but is not limited to, describing what our variables and files are used for, describing what loops are doing, describing each test, maybe including expected results and how we are manipulating the data and the many data fields. A hash mark, #, precedes each line of a comment. The script stub that follows is on this book’s companion Web site at www.wiley. com/compbooks/michael. The name is script.stub. It has all of the comments ready to get started writing a shell script. The script.stub file can be copied to a new filename. Edit the new filename, and start writing code. The script.stub file is shown in Listing 1.1.
#!/usr/bin/ksh # # SCRIPT: NAME_of_SCRIPT # AUTHOR: AUTHORS_NAME # DATE: DATE_of_CREATION # REV: 1.1.A (Valid are A, B, D, T and P) # (For Alpha, Beta, Dev, Test and Production) # # PLATFORM: (SPECIFY: AIX, HP-UX, Linux, Solaris # or Not platform dependent) # # PURPOSE: Give a clear, and if necessary, long, description of the # purpose of the shell script. This will also help you stay # focused on the task at hand. # # REV LIST: # DATE: DATE_of_REVISION # BY: AUTHOR_of_MODIFICATION # MODIFICATION: Describe what was modified, new features, etc-# # # set -n # Uncomment to check your syntax, without execution. # # NOTE: Do not forget to put the comment back in or # # the shell script will not execute! # set -x # Uncomment to debug this shell script (Korn shell only) # ########################################################## ########### DEFINE FILES AND VARIABLES HERE ############## ##########################################################

5

##########################################################

Listing 1.1 script.stub shell script starter listing. (continues)

Free & Share & Open

6

Chapter 1

############### DEFINE FUNCTIONS HERE #################### ##########################################################

########################################################## ################ BEGINNING OF MAIN ####################### ##########################################################

# End of script

Listing 1.1 script.stub shell script starter listing. (continued)

The shell script starter shown in Listing 1.1 gives you the framework to start writing the shell script with sections to declare variables and files, create functions, and write the final section, BEGINNING OF MAIN, where the main body of the shell script is written.

Control Structures
The following control structures will be used extensively. if ... then Statement if [ test_command ] then commands fi

if ... then ... else Statement if [ test_command ] then commands else commands fi

Scripting Quick Start and Review if ... then ... elif ... (else) Statement if [ test_command ]

7

then commands elif [ test_command ] then commands elif [ test_command ] then commands . . . else

(Optional) commands

fi

for ... in Statement for loop_variable in argument_list do commands done

while Statement while test_command_is_true do commands done

until Statement until do test_command_is_true

Free & Share & Open

8

Chapter 1 commands done

case Statement case $variable match_1) commands_to_execute_for_1 ;; match_2) commands_to_execute_for_2 ;; match_3) commands_to_execute_for_3 ;; . . . in

*)

(Optional - any other value) commands_to_execute_for_no_match ;;

esac

N OT E The last part of the case statement:
*) commands_to_execute_for_no_match

;; is optional.

Scripting Quick Start and Review

9

Using break, continue, exit, and return
It is sometimes necessary to break out of a for or while loop, continue in the next block of code, exit completely out of the script, or return a function’s result back to the script that called the function. break is used to terminate the execution of the entire loop, after completing the execution of all of the lines of code up to the break statement. It then steps down to the code following the end of the loop. continue is used to transfer control to the next set of code, but it continues execution of the loop. exit will do just what one would expect: It exits the entire script. An integer may be added to an exit command (for example, exit 0), which will be sent as the return code. return is used in a function to send data back, or return a result, to the calling script.

Here Document
A here document is used to redirect input into an interactive shell script or program. We can run an interactive program within a shell script without user action by supplying the required input for the interactive program, or interactive shell script. This is why it is called a here document: The required input is here, as opposed to somewhere else.

Syntax for a Here Document program_name | || & && date echo sleep wc head tail diff sdiff spell lp, lpr, enq, qprt lpstat enable disable DESCRIPTIONp Remove a file from the system Rename a file Create a directory Remove a directory Pattern matching grep command for extended regular expressions Used to locate files and directories Append to the end of a file Redirect, create, or overwrite a file Pipe, used to string commands together Logical OR—command1 || command2—execute command2 if command1 fails Execute in background Logical AND—command1 && command2—execute command2 if command1 succeeds Display the system date and time Write strings to standard output Execution halts for the specified number of seconds Count the number of words, lines, and characters in a file View the top of a file View the end of a file Compare two files Compare two files side by side (requires 132-character display) Spell checker Print a file Status of system print queues Enable, or start, a print queue Disable, or stop, a print queue (continues)

11

Free & Share & Open

12

Chapter 1
Table 1.2 Unix Commands Review (Continued) DESCRIPTIONp Display a calendar Display information about users on the system Extended who command Display $LOGNAME or $USER environment parameter Display login name, terminal, login date/time, and where logged in Information about logged-in users including the users .plan and .project Two users have a split screen conversation Display a message on a user’s screen Display a message on all logged-in users’ screens Display a message to all users on a remote host Execute a command, or log in, on a remote host Filesystems statistics Information on currently running processes Show network status Show virtual memory status Show input/output status Name of the current operating system, as well as machine information System activity report Base filename of a string parameter Display the on-line reference manual Switch to another user, also known as super-user Write out selected characters Programming language to parse characters Programming language for character substitution Start the vi editor Start the emacs editor

COMMAND cal who w whoami who am I f, finger talk write wall rwall rsh or remsh df ps netstat vmstat iostat uname sar basename man su cut awk sed vi emacs

Most of the commands shown in Table 1.2 are used at some point in this book, depending on the task we are working on in each chapter.

Scripting Quick Start and Review

13

Symbol Commands
The symbols shown in Table 1.3 are actually commands. All of the symbol commands shown in Table 1.3 are used extensively in this book.

Variables
A variable is a character string to which we assign a value. The value assigned could be a number, text, filename, device, or any other type of data. A variable is nothing more than a pointer to the actual data. We are going to use variables so much in our scripts that it will be unusual for us not to use them. In this book we are always going to specify a variable in uppercase—for example, UPPERCASE. Using uppercase variable names is not recommended in the real world of shell programming, though, because these uppercase variables may step on system environment variables, which are also in uppercase. Uppercase variables are used in this book to emphasize the variables and to make them stand out in the code. When you write your own shell scripts or modify the scripts in this book, make the variables lowercase text. To assign a variable to point to data, we use UPPERCASE=”value_to_assign” as the assignment syntax. To access the data that the variable, UPPERCASE, is pointing to, we must add a dollar sign, $, as a prefix—for example, $UPPERCASE. To view the data assigned to the variable, we use echo $UPPERCASE, print $UPPERCASE for variables, or cat $UPPERCASE, if the variable is pointing to a file, as a command structure.

Command-Line Arguments
The command-line arguments $1, $2, $3,...$9 are positional parameters, with $0 pointing to the actual command, program, shell script, or function and $1, $2, $3, ...$9 as the arguments to the command.

Table 1.3 Symbol Commands () (( )) $(( )) [] [[ ]] $( ) `command` Run the enclosed command in a sub-shell Evaluate and assign value to variable and do math in a shell Evaluate the enclosed expression Same as the test command Used for string comparison Command substitution Command substitution

Free & Share & Open

14

Chapter 1
The positional parameters, $0, $2, etc., in a function, are for the function’s use and may not be in the environment of the shell script that is calling the function. Where a variable is known in a function or shell script is called the scope of the variable.

Shift Command
The shift command is used to move positional parameters to the left; for example, shift causes $2 to become $1. We can also add a number to the shift command to move the positions more than one position; for example, shift 3 causes $4 to move to the $1 position. Sometimes we encounter situations where we have an unknown or varying number of arguments passed to a shell script or function, $1, $2, $3... (also known as positional parameters). Using the shift command is a good way of processing each positional parameter in the order they are listed. To further explain the shift command, we will show how to process an unknown number of arguments passed to the shell script shown in Listing 1.2. Try to follow through this example shell script structure. This script is using the shift command to process an unknown number of command-line arguments, or positional parameters. In this script we will refer to these as tokens.
#!/usr/bin/sh # # SCRIPT: shifting.sh # # AUTHOR: Randy Michael # # DATE: 01-22-1999 # # REV: 1.1.A # # PLATFORM: Not platform dependent # # PURPOSE: This script is used to process all of the tokens which # Are pointed to by the command-line arguments, $1, $2, $3,etc... # # REV. LIST: # # # Initialize all variables COUNT=0 # Initialize the counter to zero NUMBER=$# # Total number of command-line arguments to process # Start a while loop

while [ $COUNT -lt $NUMBER ]

Listing 1.2 Example of using the shift command.

Scripting Quick Start and Review

15

do COUNT=`expr $COUNT + 1` TOKEN=’$’$COUNT # A little math in the shell script # Loops through each token starting with $1 process each $TOKEN shift done # Grab the next token, i.e. $2 becomes $1

Listing 1.2 Example of using the shift command. (continued)

We will go through similar examples of the shift command in great detail later in the book.

Special Parameters $* and $@
There are special parameters that allow accessing all of the command-line arguments at once. $* and $@ both will act the same unless they are enclosed in double quotes, “ “.

Special Parameter Definitions
The $* special parameter specifies all command-line arguments. The $@ special parameter also specifies all command-line arguments. The “$*” special parameter takes the entire list as one argument with spaces between. The “$@” special parameter takes the entire list and separates it into separate arguments. We can rewrite the shell script shown in Listing 1.2 to process an unknown number of command-line arguments with either the $* or $@ special parameters:
#!/usr/bin/sh # # SCRIPT: shifting.sh # AUTHOR: Randy Michael # DATE: 01-22-1999 # REV: 1.1.A # PLATFORM: Not platform dependent # # PURPOSE: This script is used to process all of the tokens which

Free & Share & Open

16

Chapter 1
# Are pointed to by the command-line arguments, $1, $2, $3, etc... # # REV LIST: # # # Start a for loop

for TOKEN in $* do process each $TOKEN done

We could have also used the $@ special parameter just as easily. As we see in the previous code segment, the use of the $@ or $* is an alternative solution to the same problem, and it was less code to write. Either technique accomplishes the same task.

Double Quotes “, Forward Tics ’, and Back Tics `
How do we know which one of these to use in our scripts, functions, and command statements? This decision causes the most confusion in writing scripts. We are going to set this straight now. Depending on what the task is and the output desired, it is very important to use the correct enclosure. Failure to use these correctly will give unpredictable results. We use “, double quotes, in a statement where we want to allow character or command substitution. The “-key is located next to the Enter key on a standard USA QWERT keyboard. Use the SHIFT “-key sequence. We use ‘, forward tics, in a statement where we do not want character or command substitution. Enclosing in ‘, forward tics, is intended to use the literal text in the variable or command statement, without any substitution. All special meanings and functions are removed. It is also used when you want a variable reread each time it is used; for example, ‘$PWD‘ is used a lot in processing the PS1 command-line prompt. The ‘-key is located next to the Enter key on a standard USA QWERT keyboard. Additionally, preceding the same string with a backslash, \, also removes the special meaning of a character, or string. We use `, back tics, in a statement where we want to execute a command, or script, and have its output substituted instead; this is command substitution. The `-key is located below the Escape key, Esc, in the top-left corner of a standard USA QWERT keyboard. Command substitution is also accomplished by using the $(command) command syntax. We are going to see many different examples of these throughout this book.

Scripting Quick Start and Review

17

Math in a Shell Script
We can do arithmetic in a shell script easily. The Korn shell let command and the ((expr)) command expressions are the most commonly used methods to evaluate an integer expression. Later we will also cover the bc function to do floating-point arithmetic.

Operators
The Korn shell uses arithmetic operators from the C programming language (see Table 1.4), in decreasing order of precedence. A lot of these math operators are used in the book, but not all. In this book we try to keep things very straightforward and not confuse the reader with obscure expressions.

Table 1.4 Math Operators OPERATOR ++ — + !~ */% +> = == != & ^ | && || DESCRIPTION Auto-increment and auto-decrement, both prefix and postfix Unary plus Unary minus Logical negation; binary inversion (one’s complement) Multiplication; division; modulus (remainder) Addition; subtraction Bitwise left shift; bitwise right shift Less than or equal to; greater than or equal to Less than; greater than Equality; inequality (both evaluated left to right) Bitwise AND Bitwise exclusive OR Bitwise OR Logical AND Logical OR

Free & Share & Open

18

Chapter 1

Built-In Mathematical Functions
The Korn shell provides access to the standard set of mathematical functions. They are called using C function call syntax. Table 1.5 shows a list of shell functions. We do not have any shell scripts in this book that use any of these built-in Korn shell functions except for the int function to extract the integer portion of a floating-point number.

File Permissions, suid and sgid Programs
After writing a shell script we must remember to set the file permissions to make it executable. We use the chmod command to change the file’s mode of operation. In addition to making the script executable, it is also possible to change the mode of the file to always execute as a particular user (suid) or to always execute as a member of a particular system group (sgid). This is called setting the sticky bit. If you try to suid or sgid a shell script, it is ignored for security reasons.

Table 1.5 NAME abs log acos sin asin sinh cos sqrt cosh tan exp tanh int

Built-In Shell Functions FUNCTION Absolute value Natural logarithm Arc cosine Sine Arc sine Hyperbolic sine Cosine Square root Hyperbolic cosine Tangent Exponential function Hyperbolic tangent Integer part of floating-point number

Scripting Quick Start and Review
Setting a program to always execute as a particular user, or member of a certain group, is often used to allow all users, or a set of users, to run a program in the proper environment. As an example, most system check programs need to run as an administrative user, sometimes root. We do not want to pass out passwords so we can just make the program always execute as root and it makes everyone’s life easier. We can use the options shown in Table 1.6 in setting file permissions. Also, please review the chmod man page. By using combinations from the chmod command options, you can set the permissions on a file or directory to anything that you want. Remember that setting a shell script to suid or sgid is ignored by the system.

19

chmod Command Syntax for Each Purpose
To Make a Script Executable chmod 754 my_script.sh

or chmod u+rwx,g+rx,o+r my_script.ksh

Table 1.6 chmod Permission Options 4000 2000 1000 0400 0200 0100 0040 0020 0010 0004 0002 0001 Sets user ID on execution. Sets group ID on execution. Sets the link permission to directories or sets the save-text attribute for files. Permits read by owner. Permits write by owner. Permits execute or search by owner. Permits read by group. Permits write by group. Permits execute or search by group. Permits read by others. Permits write by others. Permits execute or search by others.

Free & Share & Open

20

Chapter 1
The owner can read, write, and execute. The group can read and execute. The world can read. To Set a Program to Always Execute as the Owner chmod 4755 my_program

The program will always execute as the owner of the file, if it is not a shell script. The owner can read, write, and execute. The group can read and execute. The world can read and execute. So no matter who executes this file it will always execute as if the owner actually executed the program. To Set a Program to Always Execute as a Member of the File Owner’s Group chmod 2755 my_program

The program will always execute as a member of the file’s group, as long as the file is not a shell script. The owner of the file can read, write, and execute. The group can read and execute. The world can read and execute. So no matter who executes this program it will always execute as a member of the file’s group. To Set a Program to Always Execute as Both the File Owner and the File Owner’s Group chmod 6755 my_program

The program will always execute as the file’s owner and as a member of the file owner’s group, as long as the program is not a shell script. The owner of the file can read, write, and execute. The group can read and execute. The world can read and execute. No matter who executes this program it will always execute as the file owner and as a member of the file owner’s group.

Running Commands on a Remote Host
We sometimes want to execute a command on a remote host and have the result displayed locally. An example would be getting filesystem statistics from a group of machines. We can do this with the rsh command. The syntax is rsh hostname command_to_execute. This is a handy little tool but two system files will need to be set up on all of the hosts before the rsh command will work. The files are .rhosts, which would be created in the user’s home directory and have the file permissions of 600, and the /etc/hosts.equiv file. For security reasons the .rhosts and hosts.equiv files, by default, are not set up to allow the execution of a remote shell. Be careful! The systems’ security could be threatened. Refer to each operating system’s documentation for details on setting up these files.

Scripting Quick Start and Review
Speaking of security, a better solution is to use Open Secure Shell (OpenSSH) instead of rsh. OpenSSH is a freeware encrypted replacement for rsh, telnet, and ftp, for the most part. To execute a command on another machine using OpenSSH use the following syntax. ssh user@hostname command_to_execute

21

This command prompts you for a password if the encryption key pairs have not been set up on both machines. Setting up the key pair relationships usually takes less than one hour. The details of the procedure are shown in the ssh man page (man ssh). The OpenSSH code can be downloaded from the following URL: www.openssh.org.

Setting Traps
When a program is terminated before it would normally end, we can catch an exit signal. This is called a trap. Table 1.7 lists some of the exit signals. To see the entire list of supported signals for your operating system, enter the following command:
# kill -l [That’s kill -(ell)]

This is a really nice tool to use in our shell scripts. On catching a trapped signal we can execute some cleanup commands before we actually exit the shell script. Commands can be executed when a signal is trapped. If the following command statement is added in a shell script, it will print to the screen “EXITING on a TRAPPED SIGNAL” and then make a clean exit on the signals 1, 2, 3, and 15. We cannot trap a kill -9. trap ‘echo “\nEXITING on a TRAPPED SIGNAL”;exit’ 1 2 3 15

We can add all sorts of commands that may be needed to clean up before exiting. As an example we may need to delete a set of files that the shell script created before we exit.
Table 1.7 Exit Signals 0 1 2 3 9 15 24 — SIGHUP SIGINT SIGQUIT SIGKILL SIGTERM SIGSTOP Normal termination, end of script Hang up, line disconnected Terminal interrupt, usually CONTROL-C Quit key, child processes to die before terminating kill -9 command, cannot trap this type of exit status kill command’s default action Stop, usually CONTROL-z

Free & Share & Open

22

Chapter 1

User Information Commands
Sometimes we need to query the system for some information about users on the system.

who Command
The who command gives this output for each logged-in user: username, tty, login time, and where the user logged in from: rmichael root pts/0 pts/1 Mar 13 10:24 Mar 15 10:43 10.10.10.6 (yogi)

w Command
The w command is really an extended who. The output looks like the following:
12:29PM User rmichael root up 27 days, 21:53,2 users, load average 1.03, 1.17, 1.09 tty login@ idle JCPU PCPU what pts/0 Mon10AM 0 3:00 1 w pts/1 10:42AM 37 5:12 5:12 tar

Notice that the top line of the preceding output is the same as the output of the uptime command. The w command gives a more detailed output than the who command by listing job process time, total user process time, but it does not reveal where the users have logged in from. We often are interested in this for security purposes. One nice thing about the w command’s output is that it also lists what the users are doing at the instant the command is entered. This can be very useful.

last Command
The last command shows the history of who has logged into the system since the wtmp file was created. This is a good tool when you need to do a little investigation of who logged into the system and when. The following is example output: root root root root root root root root root ftp pts/3 pts/2 pts/1 pts/0 pts/0 ftp ftp tty0 booboo mrranger mrranger mrranger mrranger mrranger booboo booboo Aug Aug Aug Aug Aug Aug Aug Aug Aug 06 06 06 06 06 06 06 06 06 19:22 18:45 18:45 18:44 18:44 18:43 18:19 18:18 18:06 - 19:23 still still still still - 18:44 - 18:20 - 18:18 still (00:01) logged in. logged in. logged in. logged in. (00:01) (00:00) (00:00) logged in.

Scripting Quick Start and Review root reboot shutdown root root root root tty0 ~ tty0 ftp ftp ftp ftp Aug Aug Jul Jul Jul Jul Jul 02 02 31 31 31 31 31 12:24 12:00 23:23 21:19 21:19 20:42 20:41 - 17:59 (4+05:34)

23

booboo bambam booboo bambam

-

21:19 21:19 20:42 20:42

(00:00) (00:00) (00:00) (00:00)

The output of the last command shows the username, the login port, where the user logged in from, the time of the login/logout, and the duration of the login session.

ps Command
The ps command will show information about current system processes. The ps command has many switches that will change what we look at. Some common command options are listed in Table 1.8.

Communicating with Users
Communicate with the system’s users and let them know what is going on! All Systems Administrators have the maintenance window where we can finally get control and handle some offline tasks. This is just one example of a need to communicate with the systems’ users, if any are still logged in. The most common way to get information to the system users is to use the /etc/motd file. This file is displayed each time the user logs in. If users stay logged in for days at a time they will not see any new messages of the day. This is one reason why real-time communication is needed. The commands shown in Table 1.9 allow communication to, or between, users who are currently logged in the system.

Table 1.8 Some ps Command Options ps ps -f ps -ef ps -A ps -Kf ps auxw The user’s currently running processes Full listing of the user’s currently running processes Full listing of all processes, except kernel processes All processes including kernel processes Full listing of kernel processes Wide listing sorted by percentage of CPU usage, %CPU

Free & Share & Open

24

Chapter 1
Table 1.9 wall rwall write talk Commands for Real-Time User Communication Writes a message on the screen of all logged-in users on the local host. Writes a message on the screen of all logged-in users on a remote host. Writes a message to an individual user. The user must currently be logged-in. Starts an interactive program that allows two users to have a conversation. The screen is split in two, and both users can see what each person is typing.

N OT E When using these commands be aware that if a user is using a program—for example, an accounting software package—and has that program’s screen on the terminal, then the user may not get the message or the user’s screen may become scrambled.
In addition to the preceding commands, there is a script on the Web site that accompanies this book named broadcast.ksh that can be used to send pop-up messages in a Windows (95, 98, and NT) environment. The script uses Samba, and it must be installed, and enabled, for broadcast.ksh to work. The details are in Chapter 25.

Uppercase or Lowercase Text for Easy Testing
We often need to test text strings like filenames, variables, file text, and so on, for comparison. It can vary so widely that it is easier to uppercase or lowercase the text for ease of comparison. The tr and typeset commands can be used to uppercase and lowercase text. This makes testing for things like variable input a breeze. Here is an example using the tr command: VARIABLE VALUES Expected input: Real input: Possible input: UPCASING
UPCASEVAR=$(echo $VARIABLE | tr ‘[a-z]’ ‘[A-Z]’)

TRUE TRUE true TRUE True True, etc...

DOWNCASING
DOWNCASEVAR=$(echo $VARIABLE | tr ‘[A-Z]’ ‘[a-z]’)

Scripting Quick Start and Review
In the preceding example of the tr command, we echo the string and use a pipe (|) to send the output of the echo statement to the tr command. As the preceding examples show, uppercasing uses ‘[a-z]’ ‘[A-Z]’ .

25

N OT E The single quotes are required around the square brackets.
‘[a-z]’ ‘[A-Z]’ ‘[A-Z]’ ‘[a-z]’ Used for lower to uppercase Used for upper to lowercase

No matter what the user input is, we will always have the stable input of TRUE, if uppercased, and true, if lowercased. This reduces our code testing and also helps the readability of the script. We can also use typeset to control the attributes of a variable in the Korn shell. In the previous example we are using the variable, VARIABLE. We can set the attribute to always translate all of the characters to uppercase or lowercase. To set the case attribute of VARIABLE to always translate characters to uppercase we use: typeset -u VARIABLE

The -u switch to the typeset command is used for uppercase. After we set the attribute of the variable VARIABLE, using the typeset command, any time we assign text characters to VARIABLE they are automatically translated to uppercase characters. EXAMPLE: typeset -u VARIABLE VARIABLE=”True” echo $VARIABLE TRUE

To set the case attribute of the variable VARIABLE to always translate characters to lowercase we use: typeset -l VARIABLE

EXAMPLE: typeset -l VARIABLE VARIABLE=”True” echo $VARIABLE true

Check the Return Code
Whenever we run a command there is a response back from the system about the last command that was executed, known as the return code. If the command was successful the return code will be 0, zero. If it was not successful the return will be something

Free & Share & Open

26

Chapter 1 other than 0, zero. To check the return code we look at the value of the $? shell variable. As an example, we want to check if the /usr/local/bin directory exists. Each of these blocks of code accomplishes the exact same thing: test -d /usr/local/bin if [ “$?” -eq 0 ] # Check the return code then # The return code is zero echo ‘/usr/local/bin does exist’ else # The return code is NOT zero

echo ‘/usr/local/bin fi

does NOT exist’

or if test then -d /usr/local/bin # The return code is zero does exist’

echo ‘/usr/local/bin else

# The return code is NOT zero echo ‘/usr/local/bin does NOT exist’

fi

or
If [ -d then /usr/local/bin ] # The return code is zero

echo ‘/usr/local/bin does exist’ else # The return code is NOT zero echo ‘/usr/local/bin does NOT exist’ fi

Notice that we checked the return code using $? once. The other examples use the control structure’s built-in test. The built-in tests do the same thing of processing the return code, but the built-in tests hide this step in the process. All three of the previous examples give the exact same result. This is just a matter of personal choice and readability.

Scripting Quick Start and Review

27

Time-Based Script Execution
We write a lot of shell scripts that we want to execute on a timed interval or run once at a specific time. This section addresses these needs with several examples.

Cron Tables
A cron table is a system file that is read every minute by the system and will execute any entry that is scheduled to execute in that minute. By default, any user can create a cron table with the crontab -e command, but the Systems Administrator can control which users are allowed to create and edit cron tables with the cron.allow and cron.deny files. When a user creates his or her own cron table the commands, programs, or scripts will execute in that user’s environment. It is the same thing as running the user’s $HOME/.profile before executing the command. The crontab -e command starts the default text editor, vi or emacs, on the user’s cron table.

N OT E When using the crontab command, the current user ID is the cron table that is acted on. To list the contents of the current user’s cron table, issue the crontab -l command.

Cron Table Entry Syntax
It is important to know what each field in a cron table entry is used for. Figure 1.1 shows the usage for creating a cron table entry. This cron table entry in Figure 1.1 executes the script, /usr/local/bin/ somescript.ksh, at 3:15AM, January 8, on any day of the week that January 8 falls on. Notice that we used a wildcards for the weekday field. The following cron table entry is another example:
1 0 1 1 * /usr/bin/banner “Happy New Year” > /dev/console

Minute (0 through 29) Hour (0 through 23) Day of the Month (1 through 31) Month (1 through 12) Weekday (0 - 6 for Sunday to Saturday) 15 /dev/null Figure 1.1 Cron table entry definitions and syntax. 3 8 1 * /usr/local/bin/somescript.sh 2>&1 >

Free & Share & Open

28

Chapter 1
At 1 minute after midnight on January 1, on any weekday, this cron table entry writes to the system’s console (/dev/console) Happy New Year in large banner letters.

Wildcards
* ? Match any number of characters Match a single character

at Command
Like a cron table, the at command executes commands based on time. Using the at command we can schedule a job to run once, at a specific time. When the job is executed the at command will send an e-mail, of the standard output and standard error, to the user who scheduled the job to run, unless the output is redirected. As a Systems Administrator we can control which users are allowed to schedule jobs with the at.allow and at.deny files. Refer to each operating system’s man pages before modifying these files and the many ways to use the at command for timed controlled command execution.

Output Control
How is the script going to run? Where will the output go? These questions come under job control.

Silent Running
To execute a script in silent mode we can use the following syntax:
/PATH/script_name 2>&1 > /dev/null

In this command statement the script_name shell script will execute without any output to the screen. The reason for this is that the command is terminated with the following:
2>&1 > /dev/null

By terminating a command like this it redirects standard error (stderr), specified by file descriptor 2, to standard output (stdout), specified by file descriptor 1. Then we have another redirection to /dev/null, which sends all of the output to the bit bucket. We can call this silent running. This means that there is absolutely no output from the script going to our screen. Inside the script there may be some output directed to files or devices, a particular terminal, or even the system’s console, /dev/console, but

Scripting Quick Start and Review none to the user screen. This is especially useful when executing a script from one of the system’s cron tables. In the following example cron table entry, we want to execute a script named /usr/local/bin/systemcheck.ksh, which needs to run as the root user, every 15 minutes, 24 hours a day, 7 days a week and not have any output to the screen. There will not be any screen output because we are going to end the cron table entry with:
2>&1 > /dev/null

29

Inside the script it may do some kind of notification such as paging staff or sending output to the system’s console, writing to a file or a tape device, but output such as echo “Hello world” would go to the bit bucket. But echo “Hello world” > /dev/console would go to the system’s defined console if this command statement was within the shell script. This cron table entry would need to be placed in the root cron table (must be logged in as the root user) with the following syntax.
5,20,35,50 * * * * /usr/local/bin/systemcheck.ksh 2>&1 >/dev/null

N OT E Most system check type scripts need to be in the root cron table.
Of course, a user must be logged in as root to edit root’s cron table.

The previous cron table entry would execute the /usr/local/bin/system check.ksh every 15 minutes, at 5, 20, 35, and 50 minutes, each hour, 24 hours a day, 7 days a week. It would not produce any output to the screen due to the final 2>&1 > /dev/null. Of course, the minutes selected to execute can be any. We sometimes want to spread out execution times in the cron tables so that we don’t have a lot of CPU-intensive scripts and programs starting execution at the same time.

Using getopts to Parse Command-Line Arguments
The getopts command is built in to the Korn shell. It retrieves valid command-line options specified by a single character preceded by a - (minus sign) or + (plus sign). To specify that a command switch requires an argument to the switch, it is followed by a : (colon). If the switch does not require any argument then the : should be omitted. All of the options put together are called the OptionString, and this is followed by some variable name. The argument for each switch is stored in a variable called $OPTARG. If the entire OptionString is preceded by a : (colon), then any unmatched switch option causes a ? to be loaded into the VARIABLE. The form of the command follows: getopts OptionString VARIABLE [ Argument ... ]

The easiest way to explain this is with an example. For our script we need seconds, minutes, hours, days, and a process to monitor. For each one of these we want to supply an argument—that is, -s 5 -m10 -p my_backup. In this we are specifying 5 seconds,

Free & Share & Open

30

Chapter 1
10 minutes, and the process is my_backup. Notice that there does not have to be a space between the switch and the argument. This is what makes getopts so great! The command line to set up our example looks like this:
SECS=0 MINUTES=0 HOURS=0 DAYS=0 PROCESS= # Initialize all to zero

# Initialize to null

while getopts :s:m:h:d:p: TIMED 2>/dev/null do case $TIMED in s) SECS=$OPTARG ;; m) (( MINUTES = $OPTARG * 60 )) ;; h) (( HOURS = $OPTARG * 3600 )) ;; d) (( DAYS = $OPTARG * 86400 )) ;; p) PROCESS=$OPTARG ;; \?) usage exit 1 ;; esac done (( TOTAL_SECONDS = SECONDS + MINUTES + HOURS + DAYS ))

There are a few things to note in the getopts command. The getopts command needs to be part of a while loop with a case statement within the loop for this example. On each option we specified, s, m, h, d, and p, we added a : (colon) after each switch. This tells getopts that an argument is required. The : (colon) before the OptionString list tells getopts that if an unspecified option is given, to set the $TIMED variable to the ? character. This allows us to call the usage function and exit with a return code of 1. The first thing to be careful of is that getopts does not care what arguments it receives so we have to take action if we want to exit. The last thing to note is that the first line of the while loop has redirection of standard error (file descriptor 2) to the bit bucket. Any time an unexpected argument is encountered, getopts sends a message to standard error. Because we expect this to happen, we can just ignore the messages and discard them to /dev/null. We will study getopts a lot in this book.

Making a Co-Process with Background Function
We also need to cover setting up a co-process. A co-process is a communications link between a foreground and a background process. The most common question is why is this needed? In one of the scripts we are going to call a function that will handle all of

Scripting Quick Start and Review the process monitoring for us while we do the timing control in the main script. The problem arises because we need to run this function in the background and it has an infinite loop. Within this background process-monitoring function there is an infinite loop. Without the ability to tell the loop to break out, it will continue to execute on its own after the main script, and function, is interrupted. We know what this causes—one or more defunct processes! From the main script we need a way to communicate with this loop, thus background function, to tell it to break out of the loop and exit the function cleanly when the countdown is complete and if the script is interrupted, CTRL-C. To solve this little problem we kick off our proc_watch function as a co-process, in the background. How do we do this, you ask? “Pipe it to the background” is the simplest way to put it, and that is also what it looks like, too. Look at the next example code block:
############################ function trap_exit { # Tell the co-process to break out of the loop BREAK_OUT=’Y’ print -p $BREAK_OUT # Use “print -p” to talk to the co-process } ############################ function proc_watch { # This function is started as a co-process!!! while : # Loop forever do Some Code Here read $BREAK_OUT # Do NOT need a “-p” to read! if [[ $BREAK_OUT = ‘Y’ ]] then return 0 fi done } ############################ ##### Start of Main ######## ############################ ### Set a Trap ### trap ‘trap_exit; exit 2’ 1 2 3 15 TOTAL_SECONDS=300 BREAK_OUT=’N’ proc_watch |& PW_PID=$1 # Start proc_watch as a co-process!!!! # Process ID of the last background job

31

Free & Share & Open

32

Chapter 1 until (( TOTAL_SECONDS == 0 )) do (( TOTAL_SECONDs = TOTAL_SECONDS - 1 )) sleep 1 done BREAK_OUT=’Y’ # Use “print -p” to communicate with the co-process variable print -p $BREAK_OUT kill $PW_PID exit 0 # Kill the background co-process

In this code segment we defined two functions. The trap_exit function will execute on exit signals 1, 2, 3, and 15. The other function is the proc_watch function, which is the function that we want to start as a background process. As you can see in proc_watch, it has an infinite loop. If the main script is interrupted then without a means to exit the loop, within the function, the loop alone will continue to execute! To solve this we start the proc_watch as a co-process by “piping it to the background” using pipe ampersand , |&, as a suffix. Then when we want to communicate to this co-process background function we use print -p $VARIABLE_NAME. Inside the coprocess function we just use the standard read $VARIABLE_NAME. This is the mechanism that we are going to use to break out of the loop if the main script is interrupted on a trapped signal; of course, we cannot catch a kill -9 with a trap. Try setting up the scenario described previously with a background function that has an infinite loop. Then press the CTRL-C key sequence to kill the main script, and do a ps -ef | more. You will see that the background loop is still executing! Get the PID, and do a kill -9 on that PID to kill it. Of course, if the loop’s exit criteria is ever met, the loop will exit on its own.

Catching a Delayed Command Output
Have you ever had a hard time trying to catch the output of a command that has a delayed output? This can cause a lot of frustration when you just miss it! There is a little technique that allows you to catch these delayed responses. The trick is to use an until loop. Look at the code shown here.
OUTFILE=”/tmp/outfile.out” cat /dev/null > $OUTFILE # Define the output file # Create a zero size output file

# Start an until loop to catch the delayed response until [ -s $OUTFILE ] do

Scripting Quick Start and Review delayed_output_command >> $OUTFILE done # Show the resulting output more $OUTFILE

33

This code segment first defines an output file to store the delayed output data. We start with a zero-sized file and then enter an until loop that will continue until the $OUTFILE is no longer a zero-sized file, and the until loop exits. The last step is to show the user the data that was captured from the delayed output.

Fastest Ways to Process a File Line by Line
Most shell scripts work with files, and some use a file for data input. The two fastest techniques for processing a file line by line are shown in this section. The first technique feeds a while loop from the bottom. The second technique uses file descriptors. function while_read_LINE_bottom { while read LINE do echo “$LINE” : done < $FILENAME }

The function shown in the previous code feeds the while loop from the bottom, after the done. function while_read_LINE_FD { exec 3/dev/null ;; SunOS) ping -s $HOST $PACKET_SIZE $PING_COUNT 2>/dev/null ;; *) echo “\nERROR: Unsupported Operating System - $(uname)” echo “\n\t...EXITING...\n” exit 1 esac }

Free & Share & Open

38

Chapter 1
The main body of the shell script must supply the hostname to ping. This is usually done with a while loop.

Highlighting Specific Text in a File
The technique shown here highlights specific text in a file with reverse video while displaying the entire file. To add in the reverse video piece, we have to do some command substitution within the sed statement using the tput commands. Where we specify the new_string, we will add in the control for reverse video using command substitution, one to turn highlighting on and one to turn it back off. When the command substitution is added, our sed statement will look like the following: sed s/current_string/$(tput smso)new_string$(tput rmso)/g

In our case the current_string and new_string will be the same because we only want to highlight existing text without changing it. We also want the string to be assigned to a variable as in the next command: sed s/”$STRING”/$(tput smso)”$STRING”$(tput rmso)/g

Notice the double quotes around the string variable, “$STRING”. Do not forget to add the double quotes around variables! As an experiment using command substitution, try this next command statement to highlight the machine’s host name in the /etc/hosts file on any Unix machine: cat /etc/hosts | sed s/`hostname`/$(tput smso)`hostname`$(tput rmso)/g

Keeping the Printers Printing
Keeping the printers enabled in a large shop can sometimes be overwhelming. There are two techniques to keep the printers printing. One technique is for the AIX “classic” printer subsystem, and the other is for System V printing.

AIX “Classic” Printer Subsystem
To keep AIX “classic” printer subsystem print queues running use either of the following commands. enable $(enq -AW | tail +3 | grep DOWN | awk ‘{print $1}’) 2>/dev/null

or enable $(lpstat -W | tail +3 | grep DOWN | awk ‘{print $1}’) 2>/dev/null

Scripting Quick Start and Review

39

System V Printing
To keep System V printers printing use either of the following commands. lpc enable $(lpstat -a | grep ‘not accepting’ | awk ‘{print $1}’) lpc start $( lpstat -p | grep disabled | awk ‘{print $2}’) lpc up all # Enable all printing and queuing

It is a good idea to use the root cron table to execute the appropriate command every 15 minutes or so.

Automated FTP File Transfer
You can use a here document to script an FTP file transfer. The basic idea is shown here. ftp -i -v -n wilma $HOLDFILE

Note that in the find command after the -size parameter there is a plus sign (+) preceding the file size, and there is a c added as a suffix. This combination specifies files larger than $MEG_BYTES measured in bytes, as opposed to blocks.

Free & Share & Open

40

Chapter 1

Capturing a User’s Keystrokes
In most large shops there is a need, at least occasionally, to monitor a user’s actions. You may even want to audit the keystrokes of anyone with root access to the system or other administration type accounts such as oracle. Contractors on site can pose a particular security risk. Typically when a new application comes into the environment, one or two contractors are on site for a period of time for installation, troubleshooting, and training personnel on the product. The code shown next uses the script command to capture all of the keystrokes.
TS=$(date +%m%d%y%H%M%S) THISHOST=$(hostname|cut -f1-2 -d.) LOGDIR=/usr/local/logs/script LOGFILE=${THISHOST}.${LOGNAME}.$TS touch $LOGDIR/$LOGFILE # # # # # File time stamp Host name of this machine Directory to hold the logs Creates the name of the log file Creates the actual file

# Set the command prompt export PS1=”[$LOGNAME:$THISHOST]@”’$PWD> ‘ #################### RUN IT HERE ########################## chown $LOGNAME ${LOGDIR}/${LOGFILE} # Let the user own the file during the script chmod 600 ${LOGDIR}/${LOGFILE} # Change permission to RW for the owner script ${LOGDIR}/${LOGFILE} chown root ${LOGDIR}/${LOGFILE} chmod 400 ${LOGDIR}/${LOGFILE} # Start the script monitoring session # Change the ownership to root # Set permission to read-only by root

Using the bc Utility for Floating-Point Math
On Unix machines there is a utility called bc that is an interpreter for arbitraryprecision arithmetic language. The bc command is an interactive program that provides arbitrary-precision arithmetic. You can start an interactive bc session by typing bc on the command line. Once in the session you can enter most complex arithmetic expressions as you would in a calculator. The code segment shown next creates the mathematical expression for the bc utility and then uses a here document to load the expression into bc.
# Loop through each number and build a math statement that # will add all of the numbers together. for X in $NUM_LIST do ADD=”$ADD $PLUS $X”

Scripting Quick Start and Review
PLUS=”+” done ######################################################## # Do the math here by using a here document to supply # input to the bc command. The sum of the numbers is # assigned to the SUM variable. SUM=$(bc BASE_16_NUM=8#472521 [root@yogi:/scripts]> echo $BASE_16_NUM 16#735c9

Using the printf Command
Convert a base 10 number to base 8 # printf %o 20398 47656

Free & Share & Open

42

Chapter 1
Convert a base 10 number to base 16 # printf %x 20398 4fae

Create a Menu with the select Command
There are many times when you just need to provide a menu for the end user to select from, and this is where a select statement comes in. The menu prompt is assigned to the PS3 system variable, and the select statement is used a lot like a for loop. A case statement is used to specify the action to take on each selection.
PS3=”Is today your birthday? “ echo “\n” select menu_selections in Yes No Quit do case $menu_selections in Yes) echo “\nHappy Birthday!\n” ;; No) print “\nIt is someone’s birthday today...\ Sorry it is not yours\n” ;; Quit) print “\nLater tater!\n” break ;; *) print “\nInvalid Answer...Please try again\n” ;; esac done

Notice in this code segment the use of the select statement. This looks just like a for loop with a list of possible values. Next is an embedded case statement that allows us to specify the action to take when each selection is made. The output of this simple menu is shown here with a selection of each possible answer.
./select_menu.ksh 1) 2) 3) Is Yes No Quit today your birthday? 4

Invalid Answer...Please try again

Scripting Quick Start and Review
Is today your birthday? 1 Happy Birthday! Is today your birthday? 2 It is someone’s birthday today...Sorry it is not yours Is today your birthday? 3 Later tater!

43

Sending Pop-Up Messages to Windows
When we need to get the word out quickly to the clients using Windows desktops, we can use Samba on the Unix machine to send a pop-up message. A list of the Windows machines is used in a while loop, and one by one the message is sent to each desktop that is reachable and powered on. If a message is not sent to the target Windows machine, no error is produced. We cannot guarantee that all of the messages were received. The code segment to send the message is shown here.
# Loop through each host in the $WINLIST and send the pop-up message for NODE in $WINLIST do echo “Sending to ==> $NODE” echo $MESSAGE | $SMBCLIENT -M $NODE # 1>/dev/null if (($? == 0)) then echo “Sent OK ==> $NODE” else echo “FAILED to ==> $NODE Failed” fi done

The WINLIST variable contains a list of Windows machines. The MESSAGE contains the message to send, and the SMBCLIENT variable contains the fully qualified pathname to the smbclient command.

Removing Repeated Lines in a File
The uniq command is used to report and remove repeated lines in a file. This is a valuable tool for a lot of scripting and testing. The syntax is shown here.

Free & Share & Open

44

Chapter 1
If you have a file that has repeated lines named my_list and you want to save the list without the repeated lines in a file called my_list_no_repeats, use the following command:
# uniq my_list my_list_no_repeats

If you want to see a file’s output without repeated lines use the following command:
# cat repeat_file | uniq

Removing Blank Lines from a File
The easiest way to remove blank lines from a file is to use a sed statement. The following syntax removes the blank lines.
# cat my_file | sed /^$/d

Testing for a Null Variable
Variables that have nothing assigned to them are sometimes hard to deal with. The following test will ensure that a variable is either Null or has a value assigned to it. The double quotes are very important and must be used!
VAL= # Creates a NULL variable

if [[ -z “$VAL” && “$VAL” = ‘’ ]] then echo “The VAL variable is NULL” fi

or
VAL=25 if [[ ! -z “$VAL” && “$VAL” != ‘’ ]] then echo “The VAL variable is NOT NULL” fi

Scripting Quick Start and Review

45

Directly Access the Value of the Last Positional Parameter, $#
To access the value of the $# positional parameter directly, use the following command: eval ‘$’$#

or eval \$$#

There are a lot of uses for this technique, as you will see later in this book.

Remove the Columns Heading in a Command Output
There are many instances when we want to get rid of the columns heading in a command’s output. A lot of people try to use grep -v to pattern match on something unique in the heading. A much easier and more reliable method is to use the tail command. An example is shown with the df command output.
[root:yogi]@/scripts# df -k Filesystem 1024-blocks /dev/hd4 32768 /dev/hd2 1466368 /dev/hd9var 53248 /dev/hd3 106496 /dev/hd1 4096 /proc /dev/hd10opt 655360 /dev/scripts_lv 102400 /dev/lv_temp 409600

Free %Used 15796 52% 62568 96% 8112 85% 68996 36% 3892 5% 16420 98% 24012 77% 147452 65%

Iused %Iused Mounted on 1927 12% / 44801 13% /usr 1027 8% /var 245 1% /tmp 55 6% /home - /proc 16261 10% /opt 1137 5% /scripts 29 1% /tmpfs

Now look at the same output with the column headings removed.
[root:yogi]@/scripts# df -k | tail +2 /dev/hd4 32768 15796 /dev/hd2 1466368 62568 /dev/hd9var 53248 8112 /dev/hd3 106496 68996

52% 96% 85% 36%

1927 44801 1027 245

12% 13% 8% 1%

/ /usr /var /tmp

Free & Share & Open

46

Chapter 1
/dev/hd1 /proc /dev/hd10opt /dev/scripts_lv /dev/lv_temp 4096 655360 102400 409600 3892 5% 16420 98% 24012 77% 147452 65% 55 16261 1137 29 6% /home - /proc 10% /opt 5% /scripts 1% /tmpfs

Just remember to add one to the total number of lines that you want to remove.

Arrays
The Korn shell supports one-dimensional arrays. The maximum number of array elements is 1024. When an array is defined, it is automatically dimensioned to 1024 elements. A one-dimensional array contains a sequence of array elements, which are like the boxcars connected together on a train track. An array element can be just about anything, except for another array. I know, you’re thinking that you can use an array to access an array to create two- and three-dimensional arrays. If this can be done, it is beyond the scope of this book.

Loading an Array
An array can be loaded in two ways. You can define and load the array in one step with the set -A command, or you can load the array one element at a time. Both techniques are shown here. set -A MY_ARRAY alpha beta gamma

or
X=0 # Initialize counter to zero. # Load the array with the strings alpha, beta, and gamma for ELEMENT in alpha gamma beta do MY_ARRAY[$X]=$ELEMENT ((X = X + 1)) done

The first array element is referenced by 0, not 1. To access array elements use the following syntax: echo ${MY_ARRAY[2] gamma echo ${MY_ARRAY[*] alpha beta gamma echo ${MY_ARRAY[@] alpha beta gamma # Show the third array element

# Show all array elements

# Show all array elements

Scripting Quick Start and Review echo ${#MY_ARRAY[*]} 3 echo ${#MY_ARRAY[@]} 3 echo ${MY_ARRAY} alpha # Show the total number of array elements

47

# Show the total number of array elements

# Show array element 0 (the first element)

We will use arrays in shell scripts in two chapters in this book.

Testing a String
One of the hardest things to do in a shell script is to test the user’s input from the command-line. This shell script will do the trick by using regular expressions to define the string composition.
#!/bin/ksh # # SCRIPT: test_string.ksh # AUTHOR: Randy Michael # REV: 1.0.D - Used for developement # DATE: 10/15/2002 # PLATFORM: Not Platform Dependent # # PURPOSE: This script is used to test a character # string, or variable, for its composition. # Examples include numeric, lowercase or uppercase # characters, alpha-numeric characters and IP address. # # REV LIST: # # # set -x # Uncomment to debug this script # set -n # Uncomment to verify syntax without any execution. # # REMEMBER: Put the comment back or the script will # # NOT EXECUTE! # #################################################### ############## DEFINE FUNCTIONS HERE ############### #################################################### test_string () { # This function tests a character string # Must have one argument ($1) if (( $# != 1 ))

Free & Share & Open

48

Chapter 1 then # This error would be a programming error print “ERROR: $(basename $0) requires one argument” return 1 fi # Assign arg1 to the variable --> STRING STRING=$1 # This is where the string test begins case $STRING in +([0-9]).+([0-9]).+([0-9]).+([0-9])) # Testing for an IP address - valid and invalid INVALID=FALSE # Separate the integer portions of the “IP” address # and test to ensure that nothing is greater than 255 # or it is an invalid IP address. for i in $(echo $STRING | awk -F . ‘{print $1, $2, $3, $4}’) do if (( i > 255 )) then INVALID=TRUE fi done case $INVALID in TRUE) print ‘INVALID_IP_ADDRESS’ ;; FALSE) print ‘VALID_IP_ADDRESS’ ;; esac ;; +([0-1])) # Testing for 0-1 only print ‘BINARY_OR_POSITIVE_INTEGER’ ;; +([0-7])) # Testing for 0-7 only print ‘OCTAL_OR_POSITIVE_INTEGER’ ;; +([0-9])) # Check for an integer print ‘INTEGER’ ;; +([-0-9])) # Check for a negative whole number print ‘NEGATIVE_WHOLE_NUMBER’ ;; +([0-9]|[.][0-9]))

Scripting Quick Start and Review
# Check for a positive floating point number print ‘POSITIVE_FLOATING_POINT’ ;; +(+[0-9][.][0-9])) # Check for a positive floating point number # with a + prefix print ‘POSITIVE_FLOATING_POINT’ ;; +(-[0-9][.][0-9])) # Check for a negative floating point number print ‘NEGATIVE_FLOATING_POINT’ ;; +([-.0-9])) # Check for a negative floating point number print ‘NEGATIVE_FLOATING_POINT’ ;; +([+.0-9])) # Check for a positive floating point number print ‘POSITIVE_FLOATING_POINT’ ;; +([a-f])) # Test for hexidecimal or all lowercase characters print ‘HEXIDECIMAL_OR_ALL_LOWERCASE’ ;; +([a-f]|[0-9])) # Test for hexidecimal or all lowercase characters print ‘HEXIDECIMAL_OR_ALL_LOWERCASE_ALPHANUMERIC’ ;; +([A-F])) # Test for hexidecimal or all uppercase characters print ‘HEXIDECIMAL_OR_ALL_UPPERCASE’ ;; +([A-F]|[0-9])) # Test for hexidecimal or all uppercase characters print ‘HEXIDECIMAL_OR_ALL_UPPERCASE_ALPHANUMERIC’ ;; +([a-f]|[A-F])) # Testing for hexidecimal or mixed-case characters print ‘HEXIDECIMAL_OR_MIXED_CASE’ ;; +([a-f]|[A-F]|[0-9])) # Testing for hexidecimal/alpha-numeric strings only print ‘HEXIDECIMAL_OR_MIXED_CASE_ALPHANUMERIC’ ;; +([a-z]|[A-Z]|[0-9])) # Testing for any alpha-numeric string only print ‘ALPHA-NUMERIC’ ;; +([a-z])) # Testing for all lowercase characters only print ‘ALL_LOWERCASE’ ;; +([A-Z])) # Testing for all uppercase numbers only print ‘ALL_UPPERCASE’ ;;

49

Free & Share & Open

50

Chapter 1
+([a-z]|[A-Z])) # Testing for mixed case alpha strings only print ‘MIXED_CASE’ ;; *) # None of the tests matched the string coposition print ‘INVALID_STRING_COMPOSITION’ ;; esac }

#################################################### usage () { echo “\nERROR: Please supply one character string or variable\n” echo “USAGE: $THIS_SCRIPT {character string or variable}\n” } #################################################### ############# BEGINNING OF MAIN #################### #################################################### # Query the system for the name of this shell script. # This is used for the “usage” function. THIS_SCRIPT=$(basename $0) # Check for exactly one command-line argument if (( $# != 1 )) then usage exit 1 fi # Everything looks okay if we got here. Assign the # single command-line argument to the variable “STRING” STRING=$1 # Call the “test_string” function to test the composition # of the character string stored in the $STRING variable. test_string $STRING # End of script

This is a good start but this shell script does not cover everything. Play around with it and see if you can make some improvements.

Scripting Quick Start and Review

51

Summary
This chapter is just a primer to get you started with a quick review and some little tricks and tips. In the next 24 chapters we are going to write a lot of shell scripts to solve some real-world problems. Sit back and get ready to take on the Unix world! The first thing that we are going to study is the 12 ways to process a file line by line. I have seen a lot of good and bad techniques for processing a file line by line over the last 10 years, and some have been rather inventive. The next chapter presents the 12 techniques that I have seen the most; at the end of the chapter there is a shell script that times each technique to find the fastest. Read on, and find out which one wins the race. See you in the next chapter!

Free & Share & Open

CHAPTER

2
Twelve Ways to Process a File Line by Line

Have you ever created a really slick shell script to process file data and found that you have to wait until after lunch to get the results? The script may be running so slowly because of how you are processing the file. I have come up with 12 ways to process a file line by line. Some techniques are very fast, and some make you wait for half a day. The techniques used in this chapter are measurable, and I created a shell script that will time each method so that you can see which technique suits your needs. When processing an ASCII text/data file, we are normally inside a loop of some kind. Then, as we go through the file from the top to the bottom, we process each line of text. A Korn shell script is really not meant to work on text character by character, but you can do it using various techniques. The task for this chapter is to show the lineby-line parsing techniques. We are also going to look at using file descriptors as a processing technique.

Command Syntax
First, as always, we need to go over the command syntax that we are going to use. The commands that we want to concentrate on in this chapter have to deal with while loops. When parsing a file in a while loop, we need a method to read in the entire line to a variable. The most prevalent command is read. The read command is flexible in that you can extract individual strings as well as the entire line. Speaking of line, the

53

Free & Share & Open

54

Chapter 2 line command is another alternative to grab a full line of text. Some operating systems do not support the line command. I did not find the line command on Linux or Solaris; however, the line may have been added in subsequent OS releases. In addition to the read and line, we need to look at the different ways you can use the while loop, which is the major cause of fast or slow execution times. A while loop can be used as a standalone loop in a predefined configuration; it can be used in a command pipe or with file descriptors. Each method has its own set of rules. The use of the while loop is critical to get the quickest execution times. I have seen many renditions of the proper use of a while loop, and some techniques I have seen are unique.

Using File Descriptors
Under the covers of the Unix operating system, files are referenced, copied, and moved by unique numbers known as file descriptors. You already know about three of these file descriptors:
0 - stdin 1 - stdout 2 - stderr

We have redirected output using the stdout (standard output) and stderr (standard error) in other scripts in this book. This is the first time we are going to use the stdin (standard input) file descriptor. For a short definition of each of these we can talk about the devices on the computer. Standard input usually comes into the computer from the keyboard or mouse. Standard output usually has output to the screen or to a file. Standard error is where error messages are routed by commands, programs, and scripts. We have used stderr before to send the error messages to the bit bucket, or /dev/null, and also more commonly to combine the stdout and stderr outputs together. You should remember a command like the following one: some_command 2>&1

The previous command sends all of the error messages to the same output device that standard output goes to, which is normally the terminal. We can also use other file descriptors. Valid descriptor values range from 0 to 19 on most operating systems. You have to do a lot of testing when you use the upper values to ensure that they are not reserved by the system for some reason. We will see more on using file descriptors in some of the following code listings.

Creating a Large File to Use in the Timing Test
Before I get into each method of parsing the file, I want to show you a little script you can use to create a file that has the exact number of lines that you want to process. The number of characters to create on each line can be changed by modifying the LINE_LENGTH variable in the shell script, but the default value is 80. This script also uses a while loop but this time to build a file. To create a file that has 7,500 lines, you

Twelve Ways to Process a File Line by Line add the number of lines as a parameter to the shell script name. Using the shell script in Listing 2.1, you create a 7,500-line file with the following syntax:
# mk_large_file.ksh 7500

55

The full shell script is shown in Listing 2.1.
#!/bin/ksh # # SCRIPT: mk_large_file.ksh # AUTHOR: Randy Michael # DATE: 03/15/2002 # REV: 1.2.P # # PURPOSE: This script is used to create a text file that # has a specified number of lines that is specified # on the command line. # # set -n # Uncomment to check syntax without any execution # set -x # Uncomment to debug this shell script # ################################################ # Define functions here ################################################ function usage { echo “\n...USAGE ERROR...\n” echo “\nUSAGE: $SCRIPT_NAME \n” } ################################################ # Check for the correct number of parameters ################################################ if (( $# != 1 )) # Looking for exactly one parameter then usage # Usage error was made exit 1 # Exit on a usage error fi ################################################ # Define files and variables here ################################################ LINE_LENGTH=80 OUT_FILE=/scripts/bigfile # Number of characters per line # New file to create

Listing 2.1 mk_large_file.ksh shell script listing. (continues)

Free & Share & Open

56

Chapter 2

>$OUT_FILE # Initialize to a zero-sized file SCRIPT_NAME=$(basename $0) # Extract the name of the script TOTAL_LINES=$1 # Total number of lines to create LINE_COUNT=0 # Character counter CHAR=X # Character to write to the file ################################################ # BEGINNING of MAIN ################################################ while ((LINE_COUNT < TOTAL_LINES)) # Specified by $1 do CHAR_COUNT=0 # Initialize the CHAR_COUNT to zero on every new line while ((CHAR_COUNT < LINE_LENGTH)) # Each line is fixed length do echo “${CHAR}\c” >> $OUT_FILE # Echo a single character ((CHAR_COUNT = CHAR_COUNT + 1)) # Increment the character counter done ((LINE_COUNT = LINE_COUNT + 1)) # Increment the line counter echo>>$OUT_FILE # Give a newline character done

Listing 2.1 mk_large_file.ksh shell script listing. (continued)

Each line produced by the mk_large_file.ksh script is the same length. The user specifies the total number of lines to create as a parameter to the shell script.

Twelve Methods to Parse a File Line by Line
The following paragraphs describe 12 of the parsing techniques I have commonly seen over the years. I have put them all together in one shell script separated as functions. After the functions are defined, I execute each method, or function, while timing the execution using the time command. To get accurate timing results I use a file that has 7,500 lines, where each line is the same length (we built this file using the mk_large_file.ksh shell script). A 7,500-line file is an extremely large file to be parsing line by line in a shell script, about 600 MB, but my Linux machine is so fast that I needed a large file to get the timing data greater than zero! Now it is time to look at the 12 methods to parse a file line by line. Each method uses a while statement to create a loop. The only two commands within the loop are cat $LINE, to output each line as it is read, and a no-op, specified by the : (colon) character. The thing that makes each method different is how the while loop is used.

Twelve Ways to Process a File Line by Line

57

Method 1: cat $FILENAME | while read LINE
Let’s start with the most common method that I see, which is catting a file and piping the file output to a while read loop. On each loop iteration a single line of text is read into a variable named LINE. This continuous loop will run until all of the lines in the file have been processed one at a time. The pipe is the key to the popularity of this method. It is intuitively obvious that the output from the previous command in the pipe is used as input to the next command in the pipe. As an example, if I execute the df command to list filesystem statistics and it scrolls across the screen out of view, I can use a pipe to send the output to the more command, as in the following command: df | more

When the df command is executed, the pipe stores the output in a temporary system file. Then this temporary system file is used as input to the more command, allowing me to view the df command output one page/line at a time. Our use of piping output to a while loop works the same way; the output of the cat command is used as input to the while loop and is read into the LINE variable on each loop iteration. Look at the complete function in Listing 2.2. function while_read_LINE { cat $FILENAME | while read LINE do echo “$LINE” : done }

Listing 2.2 while_read_LINE function listing.

Each of these test loops is created as a function so that we can time each method using the shell script. You could also use () C-type function definition if you wanted, as shown in Listing 2.3. while_read_LINE () { cat $FILENAME | while read LINE do echo “$LINE” : done }

Listing 2.3 Using the () declaration method function listing.

Free & Share & Open

58

Chapter 2
Whether you use the function or () technique, you get the same result. I tend to use the function method more often so that when someone edits the script they will know the block of code is a function. For beginners, the word “function” helps understanding the whole shell script a lot. The $FILENAME variable is set in the main body of the shell script. Within the while loop notice that I added the no-op (:) after the echo statement. A no-op (:) does nothing, but it always has a 0, zero, return code. I use the no-op only as a placeholder so that you can cut the function code out and paste it in one of your scripts. If you should remove the echo statement and leave the no-op, the while loop will not fail; however, the loop will not do anything either.

Method 2: while read $FILENAME from Bottom
You are now entering one of my favorite methods of parsing through a file. We still use the while read LINE syntax, but this time we feed the loop from the bottom instead of using a pipe. You will find that this is one of the fastest ways to process each line of a file. The first time you see this it looks a little unusual, but it works very well. Look at the code in Listing 2.4, and we will go over the function at the end. function while_read_LINE_bottom { while read LINE do echo “$LINE” : done < $FILENAME }

Listing 2.4 while_read_LINE_bottom function listing.

We made a few modifications to the function from Listing 2.3. The cat $FILENAME to the pipe was removed. Then we use input redirection to let us read the file from the bottom of the loop. By using the < $FILENAME notation after the done loop terminator we feed the while loop from the bottom, which greatly increases the input throughput to the loop. When we time each technique, this method will stand out at the top of the list.

Method 3: while_line_LINE_Bottom
As with the read command you can use the line command directly in a while loop using the same loop technique. In this function we use the following syntax: while line LINE

Twelve Ways to Process a File Line by Line
Whether you use this syntax in a pipe or, as in this function, feed the loop from the bottom, you can see that the line command can be used in the same manner as a read statement. Study the function in Listing 2.5 and we will go over the method at the end. function while_line_LINE_bottom { while line LINE do echo $LINE : done < $FILENAME }

59

Listing 2.5 while_line_LINE_bottom function listing.

This method is like Method 2 except that we replace read with line. You will see in our timing tests that both of these techniques may look the same, but you will be surprised at the timing difference. You will have to wait for the timing script to see the results. The function in Listing 2.5 uses the line command to assign a new line of text to the LINE variable on each loop iteration. The while loop is fed from the bottom using input redirection after the done loop terminator, done < $FILENAME. Using this input redirection technique keeps the file open for reading and is one of the fastest methods of supplying input to the loop.

Method 4: cat $FILENAME | while LINE=`line`
Now we are getting into some of the “creative” methods that I have seen in some shell scripts. Not all Unix operating systems support the line command, though. I have not found the line command in my Red Hat Linux releases, but that does not mean that it is not out there somewhere in the open-source world. Using this loop strategy replaces the read command from Listings 2.2 and 2.4 with the line command in a slightly different command structure. Look at the function in Listing 2.6, and we will see how it works at the end. function cat_while_LINE_line { cat $FILENAME | while LINE=`line` do echo “$LINE” : done }

Listing 2.6 while_read_LINE_line function listing.

Free & Share & Open

60

Chapter 2
The function in Listing 2.6 is interesting. Because we are not using the read command to assign the line of text to a variable, we need some other technique. If your machine supports the line command, then this is an option. To see if your Unix box has the line command enter the following command: which line

The response should be something like /usr/bin/line. Otherwise, you will see the $PATH list that was searched, followed by “line” not found. The line command is used to grab one whole line of text at a time. The read command does the same thing if you use only one variable with the read statement; otherwise the line of text will be broken up between the different variables used in the read statement. On each loop iteration the LINE variable is assigned a whole line of text using command substitution. This is done using the LINE=`line` command syntax. The line command is executed, and the result is assigned to the LINE variable. Of course, I could have used any variable name, for example:
MY_LINE=`line` TEXT=`line`

Please notice that the single tic marks are really back tics ( `command ` ), which are located in the top left corner of most keyboards below the ESC-key. Executing a command and assigning the output to a variable is called command substitution. Look for the timing data for this technique when you run the timing script. This extra variable assignment may have quite an effect on the timing result.

Method 5: cat $FILENAME | while line LINE
Why do the extra variable assignments when using the line command? You really do not have to. Just as the read command directly assigns a line of text to the LINE variable, the line command can do the same thing. This technique is like Method 1, but we replace the read command with the line command. Check out Listing 2.7, and we will describe the method at the end. function while_line_LINE { cat $FILENAME | while line LINE do echo “$LINE” : done }

Listing 2.7 while_line_LINE function listing.

Twelve Ways to Process a File Line by Line
In Listing 2.7 we cat the $FILENAME file and use a pipe (|) to use the cat $FILENAME output as input to the while loop. On each loop iteration the line command grabs one line from the $FILENAME file and assigns it to the LINE variable. Using a pipe in this manner does not produce very fast file processing, but it is one of the most popular methods because of its ease of use. When I see a pipe used like this, the while loop is normally used with the read command instead of the line command.

61

Method 6: while LINE=`line` from the Bottom
Again, this is one of the more obscure techniques that I have seen in any shell script. This time we are going to feed our while loop from the bottom, but this time use the line command instead of the read statement to assign the text to the LINE variable. This method is similar to the last technique, but we removed the cat $FILENAME to the pipe and instead redirect input into the loop from the bottom, after the done loop terminator. Look at the function in Listing 2.8, and we will see how it works at the end. function while_LINE_line_bottom { while LINE=`line` do echo “$LINE” : done < $FILENAME }

Listing 2.8 while_LINE_line_bottom function listing.

We use command substitution to assign the line of file text to the LINE variable as we did in the previous method. The only difference is that we are feeding the while loop from the bottom using input redirection of the $FILENAME file. You should be getting the hang of what we are doing by now. As you can see there are many ways to parse through a file, but you are going to see that not all of these techniques are very good choices. This method is one of the poorer choices. Next we are going to look at the other method of command substitution. The last two methods used the line command using the syntax LINE=`line`. We can also use the LINE=$(line) technique. Is there a speed difference?

Method 7: cat $FILENAME | while LINE=$(line)
Looks familiar? This is the same method as Method 3 except for the way we use command substitution. As I stated in the beginning, we need a rather large file to parse

Free & Share & Open

62

Chapter 2 through to get accurate timing results. When we do our timing tests we may see a difference between the two command substitution techniques. Study the function in Listing 2.9, and we will cover the function at the end. function while_LINE_line_cmdsub2 { cat $FILENAME | while LINE=$(line) do echo “$LINE” : done }

Listing 2.9 while_LINE_line_cmdsub2 function listing.

The only thing we are looking for in the function in Listing 2.9 is a timing difference between the two command substitution techniques. As each line of file text enters the loop, the line command assigns the text to the LINE variable. Let’s see how Methods 4 and 7 show up in the loop timing tests because the only difference is the assignment method.

Method 8: while LINE=$(line) from the Bottom
This method is the same technique used in Listing 2.8 except for the command substitution. In this function we are going to use the LINE=$(line) technique. We are again feeding the while loop input from the bottom, after the done loop terminator. Please review the function in Listing 2.10. function while_LINE_line_bottom_cmdsub2 { while LINE=$(line) do echo “$LINE” : done < $FILENAME }

Listing 2.10 while_LINE_line_bottom_cmdsub2 function listing.

By the look of the loop structure you might assume that this while loop is very fast executing, but you will be surprised at how slow it is. The main reason is the variable assignment, but the line command has a large effect, too.

Twelve Ways to Process a File Line by Line

63

Method 9: while read LINE Using File Descriptors
So far we have been doing some very straightforward kind of loops. Have you ever used file descriptors to parse through a file? I saved the next four functions for last. The use of file descriptors is sometimes a little hard to understand. I’m going to do my best to make this easy! Under the covers of the Unix operating system, files are referenced by file descriptors. You should already know three file descriptors right off the bat. The three that I am talking about are stdin, stdout, and stderr. Standard input, or stdin, is specified as file descriptor 0. This is usually the keyboard or mouse. Standard output, or stdout, is specified as file descriptor 1. Standard output can be your terminal screen or some kind of a file. Standard error, or stderr, is specified as file descriptor 2. Standard error is how the system and programs and scripts are able to send out or suppress error messages. You can use these file descriptors in combination with one another. I’m sure that you have seen a shell script send all output to the bit bucket, or /dev/null. Look at the following command. my_shell_script.ksh >/dev/null 2>&1

The result of the previous command is to run completely silent. In other words, there is not any external output produced. Internally the script may be reading and writing to and from files and may be sending output to a specific terminal, such as /dev/console. You may want to use this technique when you run a shell script as a cron table entry or when you just are not interested in seeing any output. In the previous example we used two file descriptors. We can also use other file descriptors to handle file input and storage. In our next four timing functions we are going to use file descriptor 0 (zero), which is standard input, and file descriptor 3. On most Unix systems valid file descriptors range from 0 to 19. In our case we are going to use file descriptor 3, but we could have just as easily used file descriptor 5. There are two steps in the method we are going to use. The first step is to close file descriptor 0 by redirecting everything to our new file descriptor 3. We use the following syntax for this step: exec 3> $TIMEFILE “\nMethod 5:” “\nfunction while_line_LINE\n” >> $TIMEFILE “function while_line_LINE”

Listing 2.15 12_ways_to_parse.ksh shell script listing. (continues)

Free & Share & Open

72

Chapter 2

time echo echo echo time echo echo echo time echo echo echo time echo echo echo time echo echo echo time echo echo echo time echo echo echo time

while_line_LINE >> $TIMEFILE “\nMethod 6:” “\nfunction while_LINE_line_bottom\n” >> $TIMEFILE “function while_LINE_line_bottom” while_LINE_line_bottom >> $TIMEFILE “\nMethod 7:” “\nfunction while_LINE_line_cmdsub2\n” >> $TIMEFILE “function while_LINE_line_cmdsub2” while_LINE_line_cmdsub2 >> $TIMEFILE “\nMethod 8:” “\nfunction while_LINE_line_bottom_cmdsub2\n” >> $TIMEFILE “function while_LINE_line_bottom_cmdsub2” while_LINE_line_bottom_cmdsub2 >> $TIMEFILE “\nMethod 9:” “\nfunction while_read_LINE_FD\n” >> $TIMEFILE “function while_read_LINE_FD” while_read_LINE_FD >> $TIMEFILE “\nMethod 10:” “\nfunction while_LINE_line_FD\n” >> $TIMEFILE “function while_LINE_line_FD” while_LINE_line_FD >> $TIMEFILE “\nMethod 11:” “\nfunction while_LINE_line_cmdsub2_FD\n” >> $TIMEFILE “function while_LINE_line_cmdsub2_FD” while_LINE_line_cmdsub2_FD >> $TIMEFILE “\nMethod 12:” “\nfunction while_line_LINE_FD\n” >> $TIMEFILE “function while_line_LINE_FD” while_line_LINE_FD >> $TIMEFILE

Listing 2.15 12_ways_to_parse.ksh shell script listing. (continued)

The shell script in Listing 2.15 first defines all of the functions that we previously covered in the Methods sections. After the functions are defined, we do a little testing of the input. We are expecting exactly one command parameter, and it should be a regular file. Look at the following code block in Listing 2.16 to see the file testing.
# Test the Input # Looking for exactly one parameter (( $# == 1 )) || usage # Does the file exist as a regular file? [[ -f $1 ]] || usage

Listing 2.16 Code to test command input.

Twelve Ways to Process a File Line by Line
The first test checks to ensure that the number of command parameters, specified by the $# operator, is exactly one. Notice that we used the double parentheses mathematical test, specified as (( math test )). Additionally, we used a logical OR, specified by ||, to execute the usage function if the number of parameters is not equal to one. We use the same type of test for the file to ensure that the file exists and the file is a regular file, as opposed to a character or block special file. When we do the test, notice that we used the double bracket test for character data, specified by [[ character test ]]. This is an important distinction to note. We again use the logical OR to execute the usage function if the return code from the test is nonzero. Now we start the actual timing tests. In doing these tests we execute the Method functions one at a time. The function’s internal while loop does the file processing, but we redirect each function’s output to a file so that we have some measurable system activity. As I stated before, the timing measurements produced by the time commands go to stderr, or file descriptor 2, which will just go to the screen by default. When this shell script executes, there are three things that go to the screen, as you will see in Listing 2.17. You can also send all of this output to a file by using the following command syntax:
12_ways_to_parse.ksh /scripts/bigfile > /tmp/timing_data.out 2>&1

73

The previous command starts with the script name, followed by the file to parse through. The output is redirected to the file /tmp/timing_data.out with stderr (file descriptor 2) redirected to stdout (file descriptor 1), specified by 2>&1. Do not forget the ampersand, &, before the 1. If the & is omitted, a file with the name 1 will be created. This is a common mistake when working with file descriptors. The placement of the stderr to stdout is important in this case. If the 2>&1 is at the end of the command, you will not get the desired result, which is all of the timing data going to a data file. In some cases the placement of the 2>&1 redirection does not matter, but it does matter here.

Timing Data for Each Method
Now all of the hard stuff has been done. We have a 7,500-line file, /scripts/ bigfile, and we have our shell script written, so let’s look at which function is the fastest in Listing 2.17.
Starting File Processing of each Method Method 1: function while_read_LINE real user sys 1m30.34s 0m35.50s 0m52.13s

Listing 2.17 Timing data for each loop method. (continues)

Free & Share & Open

74

Chapter 2

Method 2: function while_read_LINE_bottom real user sys 0m5.89s 0m5.62s 0m0.16s

Method 3: function while_line_LINE_bottom real user sys 6m53.71s 0m36.62s 6m2.03s

Method 4: function cat_while_LINE_line real user sys 7m16.87s 0m51.87s 6m8.54s

Method 5: function while_line_LINE real user sys 6m50.79s 0m36.65s 5m59.66s

Method 6: function while_LINE_line_bottom real user sys 7m20.48s 0m51.01s 6m14.57s

Method 7: function while_LINE_line_cmdsub2 real user sys 7m18.04s 0m52.01s 6m10.94s

Method 8: function while_LINE_line_bottom_cmdsub2 real 7m20.34s

Listing 2.17 Timing data for each loop method. (continued)

Twelve Ways to Process a File Line by Line

75

user sys

0m50.82s 6m14.26s

Method 9: function while_read_LINE_FD real user sys 0m5.89s 0m5.53s 0m0.28s

Method 10: function while_LINE_line_FD real user sys 8m25.35s 0m50.68s 7m15.33s

Method 11: function while_LINE_line_cmdsub2_FD real user sys 8m24.58s 0m50.04s 7m16.07s

Method 12: function while_line_LINE_FD real user sys 7m54.57s 0m35.88s 7m2.26s

Listing 2.17 Timing data for each loop method. (continued)

As you can see, all file processing loops are not created equal. Two of the methods are tied for first place. Methods 2 and 9 produce the exact same real execution time at 5.89 seconds to process a 7,500-line file. Method 1 came in second at 1 minute and 30.34 seconds. The remaining methods fall far behind, ranging from almost 7 minutes to over 8 minutes and 25.35 seconds. The sorted timing output for the real time is shown in Listing 2.18. real real real 0m5.89s 0m5.89s 1m30.34s Method 2 Method 9 Method 1

Listing 2.18 Sorted timing data by method. (continues)

Free & Share & Open

76

Chapter 2

real real real real real real real real real

6m50.79s 6m53.71s 7m16.87s 7m18.04s 7m20.34s 7m20.48s 7m54.57s 8m24.58s 8m25.35s

Method Method Method Method Method Method Method Method Method

5 3 4 7 8 6 12 11 10

Listing 2.18 Sorted timing data by method. (continued)

Let’s take a look at the code for the top three techniques. The order of appearance is Method 2, 9, and 1. function while_read_LINE_bottom { while read LINE do echo “$LINE” : done < $FILENAME }

Listing 2.19 Method 2: Tied for first place.

The method in Listing 2.19 is my favorite because it is quick and intuitive to write and understand once the input redirection is explained to the beginner. function while_read_LINE_FD { exec 3 $MAIL_FILE MAIL_LIST=”randy@my.domain.com 1234567890@mypage_somebody.net” check_filesystems # This function checks the filesystems percentage

if [ -s $MAIL_FILE ] then mail -s “Filesystem Full” $MAIL_LIST < $MAIL_FILE fi

Listing 3.1 Typical mail code segment listing.

In Listing 3.1 we see a code segment that defines the MAIL_FILE and MAIL_LIST variables that we use in the mail command. After the definitions this code segment executes the function that looks for filesystems that are over the threshold. If the threshold is exceeded, then a message is appended to the $MAIL_FILE file as shown in the following code segment:
FS=/var PERCENT=98 THISHOST=$(uname -n) echo “$THISHOST: $FS is $PERCENT” | tee -a $MAIL_FILE

This code segment is from the check_filesystems function. For my machine, this echo command statement would both display the following message to the screen and append it to the $MAIL_FILE file: yogi: /var is 98%

The hostname is yogi, the filesystem is /var, and the percentage of used space is 98 percent. Notice the tee command after the pipe (|) from the echo statement. In this case we want to display the results on the screen and send an email with the same data. The tee -a command does this double duty when you pipe the output to | tee -a $FILENAME. After the check_filesystems function finishes, we test the size of the $MAIL_FILE. If it is greater than 0 bytes in size, then we send a mail message using the mail command. The following message is sent to the randy@my.domain.com and 1234567890@mypage_somebody.net email addresses: yogi: /var is 98%

Free & Share & Open

82

Chapter 3

Problems with Outbound Mail
Before we hard-code the mail command into your shell script we need to do a little test to see if we can get the email to the destination without error. To test the functionality, add the -v switch to the mail or mailx command, as shown in Listing 3.2.
# echo “Testing: /var is 98%” > /tmp/mailfile.out # mail -v -s “Filesystem Full” randy@my.domain.com < /tmp/mailfile.out AND # mail -v -s “Filesystem Full” 1234567890@mypage_somebody.net \ < /tmp/mailfile.out

Listing 3.2 Testing the mail service using mail -v.

With the -v switch added to the mail command, all of the details of the delivery are displayed on the user’s terminal. From the delivery details we can see any errors that happen until the file is considered “sent” by the local host. If the message is not delivered to the target email address, then further investigation is needed. The next two sections look at some alternative techniques.

Create a “Bounce” Account with a .forward File
I worked at one shop where only one Unix machine in the network, other than the mail server, was allowed to send email outside of the LAN. This presented a problem for all of the other machines to get the message out when a script detected an error. The solution we used was to create a user account on the Unix machine that could send email outbound. Then we locked down this user account so no one could log in remotely. Let’s say we create a user account called bounce. In the /home/bounce directory we create a file called /home/bounce/.forward. Then in the .forward file we add the email address to which we want to forward all mail. You can add as many email addresses to this file as you want, but be aware that every single email will be forwarded to each address listed in the .forward file. On this single machine that has outside LAN mailing capability we added the user bounce to the system. Then in the /home/bounce directory we created a file called .forward that has the following entries: randy@my.domain.com 1234567890@mypage_somebody.net

Automated Event Notification
This .forward file will forward all mail received by the bounce user to the randy@ my.domain.com and 1234567890@mypage_somebody.net email addresses. This way I have an email to my desktop, and I am also notified by my text pager. On all of the other machines we have two options. The first option is to edit all of the shell scripts that send email notification and change the $MAIL_LIST variable to:
MAIL_LIST=”bounce@dino.”

83

This entry assumes that the dino host is in the same domain, specified by the period that follows the hostname dino. (dino.). An easier way is to create some entries in the aliases file for sendmail. The aliases file is usually located in /etc/aliases, but you may find it in /etc/ mail/aliases on some operating systems. The format of defining an alias is a name, username, or tag, followed by one or more email addresses. The following is an example of an aliases file: admin: bounce@dino.,randy,brad,cindy,jon,pepe

This aliases file entry creates a new alias called admin that automatically sends email to the bounce account on dino and also to randy, brad, cindy, jon, and pepe. Before these changes will take effect, we need to run the newaliases command. The sendmail -bi command works, too.

Using the sendmail Command to Send Outbound Mail
In another shop where I worked, I could not send outbound mail from any user named root. The from field had to be a valid email address that is recognized by the mail server, and root is not valid. To get around this little problem I changed the command that I used from mail to sendmail. The sendmail command allows us to add the -f switch to indicate a valid internal email address for the from field. The sendmail command is in /usr/sbin/sendmail on AIX, HP-UX, and Linux, but on SunOS the location changed to /usr/lib/sendmail. Look at the function in Listing 3.3, and we will cover the details at the end. function send_notification { if [ -s $MAIL_FILE -a “$MAILOUT” = “TRUE” ]; then case $(uname) in AIX|HP-UX|Linux) SENDMAIL=”/usr/sbin/sendmail” ;; SunOS) SENDMAIL=”/usr/lib/sendmail” ;;

Listing 3.3 send_notification function listing. (continues)

Free & Share & Open

84

Chapter 3

esac echo “\nSending email notification” $SENDMAIL -f randy@$THISHOST $MAIL_LIST < $MAIL_FILE fi }

Listing 3.3 send_notification function listing. (continued)

Notice in Listing 3.3 that we added another variable, MAILOUT. This variable is used to turn on/off the email notifications. If the $MAILOUT variable points to TRUE, and the $MAIL_FILE file is nonempty, then the email is sent. If the $MAILOUT variable does not equal the string TRUE, then the email is disabled. This is just another way to control the email notifications. In the case statement we use the output of the uname command to set the correct command path for sendmail command on the Unix platform. For AIX, HP-UX, and Linux the sendmail command path is /usr/sbin. On SunOS the sendmail path is /usr/lib. We assign the correct path to the SENDMAIL variable, and we use this variable as the command to send the mail. Once the command is defined we issue the command, as shown here:
$SENDMAIL -f randy@$THISHOST $MAIL_LIST < $MAIL_FILE

We issue the sendmail command using the -f switch and follow the switch by a valid email account name, which is randy@$THISHOST. Remember that we defined the THISHOST variable to the local machine’s hostname. The from address is followed by the list of email addresses, and the message file is used by redirecting input into the sendmail command. We can also use the following syntax: cat $MAIL_FILE | $SENDMAIL -f randy@$THISHOST $MAIL_LIST

Either sendmail statement will send the mail, if the mail server and firewall allow outgoing mail.

Dial-Out Modem Software
Many good products are on the market, both freeware and commercial, that handle large amounts of paging better than any shell script could ever do. They also have the ability to dial the modem and send the message to the provider. A list of such products is shown in Table 3.1.

Automated Event Notification
Table 3.1 Products That Handle High-Volume Paging and Modem Dialing PRODUCT DESCRIPTION FREEWARE AND SHAREWARE PRODUCTS QuickPage SMS Client HylaFAX Client/server software used to send messages to alphanumeric pagers. Command-line utility for Unix that allows you to send SMS messages to cell phones and pagers. Faxing product for Unix that allows dial-in, dialout, fax-in, fax-out, and pager notifications. COMMERCIAL PRODUCTS EtherPage TelAlert FirstPAGE Enterprise-wide alphanumeric pager software product made by MobileSys. Pager notification and interactive voice response software made by Telamon. Supports all national paging networks using IXO/TAP, made by Netcon Technologies.

85

Table 3.1 shows only a sample of the products available for paging. The nice thing about these products is the ability to dial-out on a modem. At some level in every shop there is a need to use a phone line for communications instead of the network. This gives you the ability to get the message out even if the network is having a problem.

SNMP Traps
Most large shops use an enterprise monitoring tool to monitor all of the systems from a central management console. The server software is installed on a single machine called the management station. All of the managed/monitored machines have the client software installed. This client software is an SNMP agent and uses a local MIB to define the managed objects, or management variables. These managed objects define things such as the filesystems to monitor and the trigger threshold for detecting a full filesystem. When the managed object, which in this case is a full filesystem, exceeds the set threshold, a local SNMP trap is generated and the management station captures the trap and performs the predefined action, which may be to send a text page to the System Administrator. To understand what an SNMP trap is, let’s review a short explanation of each of the pieces:

Free & Share & Open

86

Chapter 3
SNMP (Simple Network Management Protocol). SNMP is a protocol used for agent communications. The most common use for the SNMP protocol is client/server system management software. MIB (Management Information Base). Each managed machine, or agent, in an SNMP-managed network maintains a local database of information (MIB) defined to the network managed machine. An SNMP-compliant MIB contains information about the property definitions of each of the managed resources. SNMP trap. Event notification to the management server from an agent-generated event, called a trap. The server management station receives and sets objects in the MIB, and the local machine, or agent, notifies the management station of client-generated events, or traps. All of the communication between the network management server and its agents, or management clients, takes place using the Simple Network Management Protocol (SNMP). The nice thing about using an enterprise management tool is that it utilizes SNMP. With most products you can write your own shell scripts using SNMP traps. The details vary for the specific syntax for each product, but with the software installed you can have your shell scripts perform that same notifications that the enterprise management software produces. Using Tivoli Netview, EcoTools, or BMC Patrol (just to name a few) you have the ability to incorporate SNMP traps into your own shell scripts for event notifications. Please refer to the product documentation for details on creating and using SNMP traps.

Summary
This chapter is intended to give a brief overview of some techniques of getting critical information out to the system management community. This chapter mainly focused on email and some different techniques for using the mail commands. The topics discussed here form the basics for notification of system problems. You should be able to extend the list of notification techniques without much effort. If you have an enterprise management solution installed at your shop, then study the vendor documentation on using and creating SNMP traps. There are books based entirely on SNMP, and the information is just too long to cover in this book, but it is an important notification method that you need to be familiar with. If you have trouble getting the email solution to work, talk with the Network Manager to find a solution. In the next chapter we move on to look at creating progress indicators to give our users feedback on long running processes. The topics include a series of dots as the processing continues, a line that appears to rotate as processing continues, and a counter that counts down to zero.

CHAPTER

4
Progress Indicator Using a Series of Dots, a Rotating Line, or a Countdown to Zero

Giving your end users feedback that a script or program is not hung is vital on long processing jobs. We sometimes write shell scripts that take a long time to completely execute—for example, system backup scripts. A good way to keep everyone content is to have some kind of progress indicator. Just about anything can be a progress indicator as long as the end user gets the idea that job processing is continuing. In this chapter we are going to examine the following three progress indicators, which are fairly common:
■■ ■■ ■■

A series of dots A rotating line A counter counting down to zero

The dots and rotating line are more common, but the countdown method does have its place where we want to specify a timeout period. Each of these methods can be started as a separate script, as a function, or we can put the code loop directly in the background. We will cover using each of these methods.

Indicating Progress with a Series of Dots
The simplest form of progress indicator is to print a period to the screen every 5 to 20 seconds. It is simple, clean, and very easy to do. As with every script we start out with

87

Free & Share & Open

88

Chapter 4 the command syntax. All we want to do is echo a dot to the screen while continuing on the same line. echo “.\c”

The echo command prints a single dot on the screen, and the backslash c, \c, specifies a continuation on the same line without a new line or carriage return. To make a series of dots we will put this single command in a loop with some sleep time between each dot. We will use a while loop that loops forever with a 10-second sleep between printing each dot on the screen. while true do echo “.\c” sleep 10 done

If, for instance, we are running a backup script and we want to use this method to indicate progress, we would put this while loop in the background and save the process ID, PID, so that we could kill the background process when the backup script is complete. First, we will just put this while loop in the background, or we can create a function with this loop and run the function in the background. Both methods are shown in Listings 4.1 and 4.2. while true do echo “.\c” done & BG_PID=$! /usr/local/bin/my_backup.ksh kill $BG_PID

Listing 4.1 Looping in the background.

To accomplish the background loop, notice that we just put an ampersand, &, after the end of the while loop, after done. The next line uses the $! operator, which saves the PID of the last background process, BG_PID=$!. The background loop starts the dots ticking, and then we kick off the backup script, /usr/local/bin/ my_backup.ksh in the foreground. When the backup script is complete, we use the kill command to stop the dots by killing the background job, specified by kill $BG_PID. We can accomplish the same task with a function, as shown in Listing 4.2.

Progress Indicators

89

function dots { while true do echo “.\c” done } ################################## ######## Begin of Main ########### ################################## dots & BG_PID=$! /usr/local/bin/my_backup.ksh kill $BG_PID

Listing 4.2 Using a background function.

The script and function in Listing 4.2 accomplish the same task but use a background function instead of just putting the while loop in the background. We still capture the PID of the dots function, specified by $!, so we can kill the function when the backup script has completed, as we did in the previous example. We could also put the loop in a separate shell script and run the external script in the background, but this would be overkill for three lines of code.

Indicating Progress with a Rotating Line
If a series of dots is too boring, then we could use a rotating line as a progress indicator. To rotate the line we will again use the echo command, but this time we need a little more cursor control. This method requires that we display, in a series, the forward slash, /, then a hyphen, -, followed by a backslash, \, and then a pipe, |, and then repeat the process. For this character series to appear seamless we need to backspace over the last character and erase it, or overwrite it with the new character that makes the line appear to rotate. We will use a case statement inside a while loop, as shown in Listing 4.3.

Free & Share & Open

90

Chapter 4

function rotate { # PURPOSE: This function is used to give the end user some feedback that # “something” is running. It gives a line twirling in a circle. # This function is started as a background process. Assign its PID # to a variable using: # # rotate & # To start # ROTATE_PID=$! # Get the PID of the last background job # # At the end of execution just break out by killing the $ROTATE_PID # process. We also need to do a quick “cleanup” of the leftover # line of rotate output. # # FROM THE SCRIPT: # kill -9 $ROTATE_PID # echo “\b\b “ INTERVAL=1 TCOUNT=”0” # Sleep time between “twirls” # For each TCOUNT the line twirls one increment

while : # Loop forever...until this function is killed do TCOUNT=`expr $TCOUNT + 1` # Increment the TCOUNT case $TCOUNT in “1”) echo ‘-’”\b\c” sleep $INTERVAL ;; “2”) echo ‘\\’”\b\c” sleep $INTERVAL ;; “3”) echo “|\b\c” sleep $INTERVAL ;; “4”) echo “/\b\c” sleep $INTERVAL ;; *) TCOUNT=”0” ;; # Reset the TCOUNT to “0”, zero. esac done } # End of Function - rotate

Listing 4.3 Rotate function.

In the function in Listing 4.3 we first define an interval to sleep between updates. If we do not have some sleep time, then the load on the system will be noticeable. We just want to give the end user some feedback, not load the system down. At least one second

Progress Indicators is needed between screen updates. Next we start an infinite while loop and use the TCOUNT variable to control which part of the rotating line is displayed during the interval. Notice that each time that we echo a piece of the rotating line, we also back up the cursor with \b and continue on the same line with \c; both are needed. This way the next loop iteration will overwrite the previous character with a new character, and then we again back up the cursor and continue on the same line. This series of characters gives the appearance of a rotating line. We use this function just like the previous example using the dots function in Listing 4.2. We start the function in the background, save the PID of the background function using the $! operator, start our time-consuming task, and kill the background rotate function when the task is complete. We could also just put the while loop in the background without using a function. In either case, when the rotating line is killed, we need to clean up the last characters on the screen. To do the cleanup we just back up the cursor and overwrite the last character with a blank space. (See Listing 4.4.)
###################################### ########## Begin of Main ############# ###################################### rotate & ROTATE_PID=$! /usr/local/bin/my_time_consuming_task.ksh kill -9 $ROTATE_PID # Cleanup... echo “\b\b “

91

# End of Example

Listing 4.4 Example of rotate function in a shell script.

These scripts work well and execute cleanly, but do not forget to give some sleep time on each loop iteration. Now we have shown the series of dots and the rotating line methods. Another method that may sometimes be beneficial is a countdown indicator.

Creating a Countdown Indicator
There may be times when you want something to time out. If we know an approximate amount of time that we want to allow for a task to finish, we can display a countdown indicator; then, when the time is up, we can take some action. Use your imagination with this one. The process we are going to use will depend on how many digits are in

Free & Share & Open

92

Chapter 4 the current countdown, for example 0 to 9, 10 to 99, 100 to 999, and 1000 to 9999. The number of digits must be taken into account because we want a smooth transition between 1000 to 999 and 100 to 99 in the countdown, as well as other digit count changes. We also want to update the screen with a new value each second as we count down to zero. This method will again require us to control the cursor as we back up over the previous output and overwrite the characters with a new countdown number. Other than the cursor control this script is not very difficult. Let’s look at the script and explain the process afterward (see Listing 4.5).

#!/bin/ksh # # SCRIPT: countdown.ksh # # AUTHOR: Randy Michael - Systems Administrator # DATE: 02-29-2000 # PLATFORM: Not Platform Dependent # # PURPOSE: This script will do the same thing as a sleep command # while giving the user feedback as to the number of seconds # remaining. It takes input between 1 and 9999 seconds only. # SCRIPT_NAME=$(basename $0) ########################################## ######## DEFINE FUNCTIONS HERE ########### ########################################## usage () { echo “\nUSAGE: $SCRIPT_NAME seconds\n” } ########################################## trap_exit () { echo “\n\n...EXITING on a trapped signal...\n” } ########################################## test_string () { # This function tests for a positive integer! if (( $# != 1 )) then

Listing 4.5 countdown.ksh shell script.

Progress Indicators

93

print ‘ERROR’ break fi STRING=$1 case $STRING in +([0-9])) print ‘POS_INT’ ;; *) print ‘NOT’ ;; esac } ########################################## ########## START OF MAIN ################# ########################################## trap ‘trap_exit;exit 2’ 1 2 3 15 if (( $# != 1 )) then usage exit 1 fi # Test for a positive integer INT_STRING=$(test_string $1) if [[ $INT_STRING != ‘POS_INT’ ]] then echo “\nINVALID INPUT ==> $1 ...EXITING...\n” usage exit 1 fi # Check for a valid range 1 - 9999 if (( $1 > 0 && $1 < 10000 )) then S=$1 # Total second to start the countdown from $S\c”

echo “Seconds Remaining:

while (( S > 0 )) # Start the loop do # In this loop we back over the previous countdown value

Listing 4.5 countdown.ksh shell script. (continues)

Free & Share & Open

94

Chapter 4

# and update the screen with a new countdown value. It # depends on how many digits the number has to determine # how many spaces to back up. sleep 1 if (( S < 10 )) then echo “\b\b # For numbers 0-9 \b\c”

elif (( S >= 10 && S < 100 )) # For numbers 10-99 then echo “\b\b\b \b\b\c” elif (( S >= 100 && S < 1000 )) # For numbers 100-999 then echo “\b\b\b\b \b\b\b\c” elif (( S >= 1000 && S < 10000 )) # For numbers 1000-9999 then echo “\b\b\b\b\b \b\b\b\b\c” fi ((S = S - 1)) # Decrement the counter by 1 echo “$S\c” done echo “\n” # Done - give a new line... else echo “Invalid input ==> $1” echo “Range 1 - 9999 seconds” usage exit 1 fi # Update the screen with the new value

Listing 4.5 countdown.ksh shell script. (continued)

Let’s review the countdown.ksh shell script in Listing 4.5 from the top. We start the script by defining the shell script’s filename. We use the basename $0 command, which will remove the leading directory path and leave only the filename. We need the script’s filename for the usage function, and we never want to hard-code a filename because we may rename the script at some point. Next, we define all of our functions. As always, we have our usage function for incorrect command-line usage. The usage function is where we need the shell script filename that we captured with the preceding basename $0 command. If the basename command were executed in the usage

Progress Indicators function the result would be usage instead of countdown.ksh. This subtle difference in using the basename command is a common mistake. Next we have the trap_exit function that will execute on trapped exit signals 1, 2, 3, and 15 (of course, we cannot trap kill -9). This trap_exit function will display ...EXITING on a trapped signal... as an informational message to the user. The test_string function is used to test for an integer value greater than or equal to 0, zero. To test for an integer we just use the regular expression +([0-9]) in a case statement. This regular expression will be true if the value is an integer value greater than or equal to 0, zero. In Chapter 1 there is a very extensive test_string.ksh shellscript that includes lowercase and uppercase characters, mixed-case strings, and numeric and alphanumeric characters. Regular expressions are great for string tests and are flexible to use. We start the main part of the script by setting a trap to catch exit signals 1, 2, 3, and 15. On these exit signals we execute our trap_exit function that we previously covered. After setting the trap we check to confirm that we have exactly one commandline argument. If we have more or less than one argument, then we run our usage function and exit with a return code of 1. The integer test for the command-line argument is next. To make this test we use our test_string function and assign the output to the variable INT_STRING. The test_string should return POS_INT, or we inform the user of the invalid value, run the usage function, and exit the script with a return code of 1. If we have got this far we know that we have a positive integer, so we need to make sure that the integer is within the valid range for this shell script. The valid range is 1 to 9999 seconds, which is 2.78 hours. If the value is out of range, then we inform the user that the value is out of range, run the usage function, and exit the script with a return code of 1. All usage errors exit with a return code of 1 in this shell script. Now we are ready to start the countdown. The countdown takes place in a while loop. Within this while loop notice the if..then..elif..elif.. control structure and the cursor control. This cursor control is dependent on the number of digits in the current countdown value. We need to control the cursor using this method so that we get a smooth transition between 1000 and 999, 100 to 99, and 10 to 9. If you do not handle the transition by cursor control the digit set will move across the screen during the transitions. For the cursor control we use the echo command with a backslash b, \b, to back the cursor one space. For three spaces we use \b\b\b\c with the final \c keeping the cursor on the same line without a new line and carriage return. So, in each loop iteration the cursor is controlled depending on the current number of digits in the current countdown value. When the countdown reaches 0, the script will output one new line and carriage return and exit with a return code of 0.

95

Other Options to Consider
As with any script, we may be able to improve on the techniques. The series-of-dots method is so simple that I cannot think of any real improvements. The rotating line is a fun little script to play with, and I have accomplished the same result in several different ways. Each method I used produced a noticeable load on the system if the sleep

Free & Share & Open

96

Chapter 4 statement was removed, so that the line twirled as fast as possible. Try to see if you can find a technique that will not produce a noticeable load and does not require a sleep of at least one second, using a shell script! In the countdown indicator the actual countdown time may not be exactly accurate. The inaccuracy is due to the variation in response time due to the load on the system. If your system is not under any load, the countdown time will be fairly stable and accurate. If you have a normally very active system, your countdown time can vary widely depending on the load and the duration of the countdown—the longer the countdown time, the less accurate the timing. A more accurate way to handle an exact timing is to use an at command to kick off the job at a specific time in the future. The following at command example will execute a script called time_out.ksh in 500 seconds: echo time_out.ksh | at now + 500 seconds

The at command is very flexible and very accurate for timing purposes. Another option is to use the shell variable SECONDS. This variable is extremely accurate and easy to use. The first step is to initialize the SECONDS variable to 0, zero. Once the variable is initialized you need only test the variable, which keeps track of the number of seconds since the SECONDS variable was initialized. Type the following lines in on the command line.
# SECONDS=0 (Wait 5 seconds...) # echo $SECONDS 5

Play around with each of these techniques, and always strive to keep your end users informed. A blank or “frozen” screen makes people uncomfortable.

Summary
In this chapter we presented three techniques to help keep our script users content. Each technique has its place, and they are all easy to implement within any shell script or function. We covered how to save the PID of the last background job and how to put an entire loop in the background. The background looping can make a script a little easier to follow if you are not yet proficient at creating and using functions. Remember, informed users are happy users! In the next chapter we will cover monitoring a system for full filesystems. Methods covered include a typical percentage method to the number of megabytes free, for very large filesystems. Chapter 5 ends with a shell script that does auto detection using the filesystem size to set the monitoring method.

CHAPTER

5
File System Monitoring

The most common monitoring task is monitoring for full filesystems. On different flavors of Unix the monitoring techniques are the same, but the commands and fields in the output vary slightly. This difference is due to the fact that command syntax and the output columns vary depending on the Unix system. We are going to step through the entire process of building a script to monitor filesystem usage and show the philosophy behind the techniques used. In scripting this solution we will cover five monitoring techniques, starting with the most basic monitoring—percentage of space used in each filesystem. The next part will build on this original base code and add exceptions capability allowing an override of the script’s set threshold for a filesystem to be considered full. The third part will deal with large filesystems, which is typically considered to be a filesystem larger than 2 gigabytes, 2GB. This script modification will use the megabytes, MB, of free space technique. The fourth part will add exception capability to the MB of free space method. The fifth part in this series combines both the percentage of used space and MB of free space techniques with an added auto-detect feature to decide how to monitor each filesystem. Regular filesystems will be monitored with percent used and large filesystems as MB of free space, and, of course, with the exception capability. The sixth and final script will allow the filesystem monitor script to run on AIX, Linux, HP-UX, or Solaris without any further modification.

97

Free & Share & Open

98

Chapter 5

In This Chapter
In this chapter, we will cover the following six shell scripts related to filesystem monitoring:
■ ■ ■ ■ ■ ■ ■ ■ ■ ■

Percentage of used space method Percentage of used space with exceptions capability Megabytes of free space method Megabytes of free space with exceptions capability Combining percentage used and megabytes of free space with exceptions capability Enabling the combined script to execute on AIX, HP-UX, Linux, and Solaris

■ ■

Syntax
Our first task, as usual, is to get the required command syntax. For this initial example we are going to monitor an AIX system (HP-UX, Linux, and Solaris will be covered later). The command syntax to look at the filesystems in kilobytes, KB, or 1024-byte blocks, is df -k in AIX. Let’s take a look at the output of the df -k command on an AIX 5L machine:
Filesystem 1024-blocks /dev/hd4 32768 /dev/hd2 1212416 /dev/hd9var 53248 /dev/hd3 106496 /dev/hd1 4096 /proc /dev/hd10opt 638976 /dev/scripts_lv 102400 /dev/cd0 656756 Free %Used Iused %Iused Mounted on 16376 51% 1663 11% / 57592 96% 36386 13% /usr 30824 43% 540 5% /var 99932 7% 135 1% /tmp 3916 5% 25 3% /home - /proc 24456 97% 15457 10% /opt 95264 7% 435 2% /scripts 0 100% 328378 100% /cdrom

The fields in the command output that we are concerned about are column 1, the Filesystem device, column 4, the %Used, and Mounted on in column 7. There are at least two reasons that we want both the filesystem device and the mount point. The first reason is to know if it is an NFS mounted filesystem. This first column will show the NFS server name as part of the device definition if it is NFS mounted. The second reason is that we will not want to monitor a mounted CD-ROM. A CD-ROM will always show that it is 100 percent used because it is mounted as read-only and you cannot write to it (I know, CD-RW drives, but these are still not the norm in business environments). As you can see in the bottom row of the preceding output, the /cdrom mount point does indeed show that it is 100 percent utilized. We want to omit this from the output

File System Monitoring along with the column heading at the top line. The first step is to show everything except for the column headings. We can use the following syntax: df -k | tail +2

99

This delivers the following output without the column headings:
/dev/hd4 /dev/hd2 /dev/hd9var /dev/hd3 /dev/hd1 /proc /dev/hd10opt /dev/scripts_lv /dev/cd0 32768 1212416 53248 106496 4096 638976 102400 656756 16376 51% 1663 11% / 57592 96% 36386 13% /usr 30824 43% 540 5% /var 99932 7% 135 1% /tmp 3916 5% 25 3% /home - /proc 24456 97% 15457 10% /opt 95264 7% 435 2% /scripts 0 100% 328378 100% /cdrom

This output looks a bit better, but we still have a couple of things we are not interested in. The /cdrom is at 100 percent all of the time, and the /proc mount point has no values, just hyphens. The /proc filesystem is new to AIX 5L, and because it has no values, we want to eliminate it from our output. Notice the device, in column 1, for the CD-ROM is /dev/cd0. This is what we want to use as a tag to pattern match on instead of the mount point because it may at some point be mounted somewhere else, for example /mnt. We may also have devices /dev/cd1 and /dev/cd2, too, if not now perhaps in the future. This, too, is easy to take care of, though. We can expand on our command statement to exclude both lines from the output with one egrep statement, as in the following: df -k | tail +2 | egrep -v ‘/dev/cd[0-9]|/proc’

In this statement we used the egrep command with a -v switch. The -v switch means to show everything except what it patterned matched on. The egrep is used for extended regular expressions; in this case, we want to exclude two rows of output. To save an extra grep statement we use egrep and enclose what we are pattern matching on within single tic marks, ‘ ‘, and separate each item in the list with a pipe symbol, |. The following two commands are equivalent: df -k | tail +2 | grep -v ‘/dev/cd[0-9]’ | grep -v ‘/proc’ df -k | tail +2 | egrep -v ‘/dev/cd[0-9]|/proc’

Also notice in both statements the pattern match on the CD-ROM devices. The grep and egrep statements will match devices /dev/cd0 up through the last device, for example /dev/cd24, using /dev/cd[0-9] as the pattern match. Do not forget the tic marks around ‘/dev/cd[0-9]’ or the grep/egrep statement may fail.

Free & Share & Open

100

Chapter 5
Using egrep saves a little bit of code, but both commands produce the same output, shown here:
/dev/hd4 /dev/hd2 /dev/hd9var /dev/hd3 /dev/hd1 /dev/hd10opt /dev/scripts_lv 32768 1212416 53248 106496 4096 638976 102400 16376 57592 30864 99932 3916 24456 95264 51% 96% 43% 7% 5% 97% 7% 1663 36386 539 134 25 15457 435 11% / 13% /usr 5% /var 1% /tmp 3% /home 10% /opt 2% /scripts

In this output we have all of the rows of data we are looking for; however, we have some extra columns that we are not interested in. Now let’s extract out the columns of interest, 1, 4, and 7. Extracting the columns is easy to do with an awk statement. Using an awk statement is the cleanest method, and the columns are selected using the positional parameters, or columns, $1, $2, $3,...,$n. As we keep building this command statement we add in the awk part of the command. df -k | tail +2 | egrep -v ‘/dev/cd[0-9]|/proc’ \ | awk ‘{print $1, $4, $7}’

First, notice that we extended our command onto the next line with the backslash character, \. This convention helps with the readability of the script. In the awk part of the statement we placed a comma and a space after each field, or positional parameter. The comma and space are needed to ensure that the fields remain separated by at least one space. This command statement leaves the following output:
/dev/hd4 51% / /dev/hd2 96% /usr /dev/hd9var 43% /var /dev/hd3 7% /tmp /dev/hd1 5% /home /dev/hd10opt 97% /opt /dev/scripts_lv 7% /scripts

For ease of working with our command output we can write it to a file and work with the file. In our script we can define a file and point to the file with a variable. The following code will work:
WORKFILE=”/tmp/df.work” >$WORKFILE # df output work file # Initialize the file to zero size

Before we go any further we also need to decide on a trigger threshold for when a filesystem is considered full, and we want to define a variable for this, too. For our example we will say that anything over 85 percent is considered a full filesystem, and we will assign this value to the variable FSMAX:
FSMAX=”85”

File System Monitoring
From these definitions we are saying that any monitored filesystem that has used more than 85 percent of its capacity is considered full. Our next step is to loop through each row of data in our output file. Our working data file is /tmp/df.work, which is pointed to by the $WORKFILE variable, and we want to compare the second column, the percentage used for each filesystem, to the $FSMAX variable, which we initialized to 85. But we still have a problem; the $WORKFILE entry still has a %, percent sign, and we need an integer value to compare to the $FSMAX value. We will take care of this conversion with a sed statement. We use sed for character substitution and, in this case, character removal. The sed statement is just before the numerical comparison in a loop that follows. Please study Listing 5.1, and pay close attention to the bold text.
#!/usr/bin/ksh # # SCRIPT: fs_mon_AIX.ksh # AUTHOR: Randy Michael # DATE: 08-22-2001 # REV: 1.1.P # PURPOSE: This script is used to monitor for full filesystems, # which is defined as “exceeding” the FSMAX value. # A message is displayed for all “full” filesystems. # # REV LIST: # # set -n # Uncomment to check syntax without any execution # set -x # Uncomment to debug this script # ##### DEFINE FILES AND VARIABLES HERE #### FSMAX=”85” # Max. FS percentage value

101

WORKFILE=”/tmp/df.work” # Holds filesystem data >$WORKFILE # Initialize to empty OUTFILE=”/tmp/df.outfile” # Output display file >$OUTFILE # Initialize to empty THISHOST=`hostname` # Hostname of this machine ######## START OF MAIN ############# # Get the data of interest by stripping out /dev/cd#, # /proc rows and keeping columns 1, 4 and 7 df -k | tail +2 | egrep -v ‘/dev/cd[0-9] | /proc’ \ | awk ‘{print $1, $4, $7}’ > $WORKFILE # Loop through each line of the file and compare column 2 while read FSDEVICE FSVALUE FSMOUNT

Listing 5.1 fs_mon_AIX.ksh shell script. (continues)

Free & Share & Open

102

Chapter 5

do FSVALUE=$(echo $FSVALUE | sed s/\%//g) # Remove the % sign typeset -i FSVALUE if [ $FSVALUE -gt $FSMAX ] then echo “$FSDEVICE mounted on $FSMOUNT is ${FSVALUE}%” \ >> $OUTFILE fi done < $WORKFILE # Feed the while loop from the bottom!! if [[ -s $OUTFILE ]] then echo “\nFull Filesystem(s) on $THISHOST\n” cat $OUTFILE print fi

Listing 5.1 fs_mon_AIX.ksh shell script. (continued)

The items highlighted in the script are all important to note. We start with getting the hostname of the machine. We want to know which machine the report is relating to. Next we load the $WORKFILE with the filesystem data. Just before the numerical test is made we remove the % sign and then typeset the variable, FSVALUE, to be an integer. Then we make the over-limit test, and if the filesystem in the current loop iteration has exceeded the threshold of 85 percent, we append a message to the $OUTFILE. Notice that the while loop is getting its data from the bottom of the loop, after done. This is the fastest technique to process a file line by line. After processing the entire file we test to see if the $OUTFILE exists and is greater than zero bytes in size. If it has data, then we print an output header, with a newline before and after, and display the $OUTFILE file followed by another blank line. In Listing 5.1 we used an assortment of commands to accomplish the same task in a different way—for example, using VARIABLE=$(command) and VARIABLE=`command`, to execute a command and assign the command’s output to a variable, and the use of the echo and print commands. In both instances the result is the same. We again see there is not just one way to accomplish the same task. We also want to explain how we use sed for character substitution. The basic syntax of the sed statement that we are going to use is as follows: command | sed s/current_string/new_string/g

When we extend our command and pipe the last pipe’s output to the sed statement we get the following: df -k | tail +2 | egrep -v ‘/dev/cd[0-9]|/proc’ \ | awk ‘{print $1, $4, $7}’ | sed s/\%//g

File System Monitoring
The point to notice about the preceding sed part of the command statement is that we had to escape the %, percent sign, with a \, backslash. This is because % is a special character in Unix. To remove the special meaning from, or to escape, the function we use a backslash before the % sign, \%. This lets us literally use % as a text character as opposed to its system-defined value or function. See Listing 5.2.
Full Filesystem(s) on yogi /dev/hd2 mounted on /usr is 96% /dev/hd10opt mounted on /opt is 97%

103

Listing 5.2 Full filesystem script in action.

This script is okay, but we really are not very concerned about these filesystems being at these current values. The reason is that /usr and /opt, on AIX, should remain static in size. The reason is that /usr is where the OS and application code for the system resides, and /opt, new to AIX 5L as a mount point, is where Linux code resides. So how can we give an exception to these two filesystems?

Adding Exceptions Capability to Monitoring
The fs_mon.ksh script is great for what it is written for, but in the real world we always have to make exceptions and we always strive to cover all of the gotchas when writing shell scripts. Now we are going to add the capability to override the default FSMAX threshold. Because we are going to be able to override the default, it would be really nice to be able to either raise or lower the threshold for individual filesystems. To accomplish this script tailoring, we need a data file to hold our exceptions. We want to use a data file so that people are not editing the shell script every time a filesystem threshold is to be changed. To make it simple, let’s use the file /usr/local/ bin/exceptions and point to the file with the EXCEPTIONS variable. Now that we know the name of the file, we need a format for the data in the $EXCEPTIONS file. A good format for this data file is the /mount_point and a NEW_MAX%. We will also want to ignore any entry that is commented out with a pound sign, #. This may sound like a lot, but it is really not too difficult to modify the script code and add a function to read the exceptions file. Now we can set it up.

The Exceptions File
To set up our exceptions file we can always use /usr/local/bin, or your favorite place, as a bin directory. To keep things nice we can define a bin directory for the script to use. This is a good thing to do in case the files need to be moved for some reason. The declarations are shown here:

Free & Share & Open

104

Chapter 5
BINDIR=”/usr/local/bin” EXCEPTIONS=”${BINDIR}/exceptions”

Notice the curly braces around the BINDIR variable when it is used to define the EXCEPTIONS file. This is always a good thing to do if the variable name will have a character, which is not associated with the variable’s name, next to the variable name without a space. Otherwise, an error may occur that could be very hard to find!
EXCEPTIONS=”$BINDIR/exceptions”

versus
EXCEPTIONS=”${BINDIR}/exceptions”

In all of the ways there are to set up exceptions capability, grep seems to come up the most. Please avoid the grep mistake! The two fields in the $EXCEPTIONS file are the /mount_point and the NEW_MAX% value. The first instinct is to grep on the /mount_point, but what if /mount_point is root, /? If you grep on /, and the / entry is not the first entry in the exceptions file, then you will get a pattern match on the wrong entry, and thus use the wrong $NEW_MAX% in deciding if the / mount point is full. In fact, if you grep on / in the exceptions file, you will get a match on the first entry in the file every time. Listing 5.3 shows some wrong code that made this very grep mistake: while read FSDEVICE FSVALUE FSMOUNT do # Strip out the % sign if it exists FSVALUE=$(echo $FSVALUE | sed s/\%//g) # Remove the % sign if [[ -s $EXCEPTIONS ]] # Do we have a non-empty file? then # Found it! # Look for the current $FSMOUNT value in the file #WRONG CODE, DON’T MAKE THIS MISTAKE USING grep!! cat $EXCEPTIONS | grep -v “^#” | grep $FSMOUNT \ | read FSNAME NEW_MAX if [ $? -eq 0 ] # Found it! then if [[ $FSNAME = $FSMOUNT ]] # Sanity check then NEW_MAX=$(echo $NEW_MAX | sed s/\%//g) if [ $FSVALUE -gt $NEW_MAX ] # Use the new $NEW_MAX then echo “$FSDEVICE mount on $FSMOUNT is ${FSVALUE}%” \ >> $OUTFILE fi elif [ $FSVALUE -gt $FSMAX ] # Not in $EXCEPTIONS file then

Listing 5.3 The wrong way to use grep.

File System Monitoring

105

echo “$FSDEVICE mount on $FSMOUNT is ${FSVALUE}%” \ >> $OUTFILE fi fi else # No exceptions file...use script default if [ $FSVALUE -gt $FSMAX ] then echo “$FSDEVICE mount on $FSMOUNT is ${FSVALUE}%” \ >> $OUTFILE fi fi done < $WORKFILE

Listing 5.3 The wrong way to use grep. (continued)

The code in Listing 5.3 really looks as if it should work, and it does some of the time! To get around the error that grep introduces, we need to just set up a function that will look for an exact match for each entry in the exceptions file. Now let’s look at this new technique. We want to write two functions, one to load the $EXCEPTIONS file data without the comment lines, the lines beginning with a #, while omitting all blank lines into a data file, and one to search through the exceptions file data and perform the tests. This is a simple one-line function to load the $EXCEPTIONS file data into the $DATA_EXCEPTIONS file: function load_EXCEPTIONS_file { # Ignore any line that begins with a pound sign, # # and also remove all blank lines cat $EXCEPTIONS | } grep -v “^#” | sed /^$/d > $DATA_EXCEPTIONS

In the preceding function we use the ^, caret character, along with the grep -v to ignore any line beginning with a #, pound sign. We also use the ^$ with the sed statement to remove any blank lines and then redirect output to a data file, which is pointed to by the $DATA_EXCEPTIONS variable. After we have the exceptions file data loaded, we have the following check_exceptions function that will look in the $DATA_EXCEPTIONS file for the current mount point and, if found, will check the $NEW_MAX value to the system’s reported percent used value. The function will present back to the script a return code relating to the result of the test. function check_exceptions { # set -x # Uncomment to debug this function

Free & Share & Open

106

Chapter 5 while read FSNAME NEW_MAX # Feeding data from Bottom of Loop!!! do if [[ $FSNAME = $FSMOUNT ]] # Correct /mount_point? then # Get rid of the % sign, if it exists! NEW_MAX=$(echo $NEW_MAX | sed s/\%//g) if [ $FSVALUE -gt $NEW_MAX ] then # Over Limit...Return a “0”, zero return 0 # FOUND OVER LIMIT - Return 0 else # Found in the file but is within limits return 2 # Found OK fi fi done < $DATA_EXCEPTIONS # Feed from the bottom of the loop!! return 1 # Not found in File }

This check_exceptions function is called during each loop iteration in the main script and returns a 0, zero, if the /mount_point is found to exceed the NEW_MAX%. It will return a 2 if the mount point was found to be OK in the exceptions data file and return a 1, one, if the mount point was not found in the $DATA_EXCEPTIONS file. There are plenty of comments throughout this new script, so feel free to follow through and pick up a few pointers—pay particular attention to the bold text in Listing 5.4.

#!/usr/bin/ksh # # SCRIPT: fs_mon_AIX_excep.ksh # AUTHOR: Randy Michael # DATE: 08-22-2001 # REV: 2.1.P # # PURPOSE: This script is used to monitor for full filesystems, # which is defined as “exceeding” the FSMAX value. # A message is displayed for all “full” filesystems. # # PLATFORM: AIX # # REV LIST: # 08-23-2001 - Randy Michael # Added code to override the default FSMAX script threshold # using an “exceptions” file, defined by the $EXCEPTIONS # variable, that list /mount_point and NEW_MAX% # # set -n # Uncomment to check syntax without any execution # set -x # Uncomment to debug this script

Listing 5.4 fs_mon_AIX_except.ksh shell script.

File System Monitoring

107

# ##### DEFINE FILES AND VARIABLES HERE #### FSMAX=”85” # Max. FS percentage value

WORKFILE=”/tmp/df.work” # Holds filesystem data >$WORKFILE # Initialize to empty OUTFILE=”/tmp/df.outfile” # Output display file >$OUTFILE # Initialize to empty BINDIR=”/usr/local/bin” # Local bin directory THISHOST=`hostname` # Hostname of this machine EXCEPTIONS=”${BINDIR}/exceptions” # Overrides $FSMAX DATA_EXCEPTIONS=”/tmp/dfdata.out” # Exceptions file w/o #, comments

####### DEFINE FUNCTIONS HERE ##### function load_EXCEPTIONS_file { # Ignore any line that begins with a pound sign, # # and omit all blank lines cat $EXCEPTIONS | } grep -v “^#” | sed /^$/d > $DATA_EXCEPTIONS

################################### function check_exceptions { # set -x # Uncomment to debug this function while read FSNAME NEW_MAX # Feeding data from Bottom of Loop!!! do if [[ $FSNAME = $FSMOUNT ]] # Correct /mount_point? then # Get rid of the % sign, if it exists! NEW_MAX=$(echo $NEW_MAX | sed s/\%//g) if [ $FSVALUE -gt $NEW_MAX ] then # Over Limit...Return a “0”, zero return 0 # FOUND OUT OF LIMITS - Return 0 fi fi done < $DATA_EXCEPTIONS # Feed from the bottom of the loop!! return 1 # Not found in File }

Listing 5.4 fs_mon_AIX_except.ksh shell script. (continues)

Free & Share & Open

108

Chapter 5

############################################# ######## START OF MAIN ############# #################################### # If there is an exceptions file...load it... [[ -s $EXCEPTIONS ]] && load_EXCEPTIONS_file # Get the data of interest by stripping out /dev/cd#, # /proc rows and keeping columns 1, 4, and 7 df -k | tail +2 | egrep -v ‘/dev/cd[0-9]|/proc’ \ | awk ‘{print $1, $4, $7}’ > $WORKFILE # Loop through each line of the file and compare column 2 while read FSDEVICE FSVALUE FSMOUNT do # Feeding the while loop from the BOTTOM!! # Strip out the % sign if it exists FSVALUE=$(echo $FSVALUE | sed s/\%//g) # Remove the % sign if [[ -s $EXCEPTIONS ]] # Do we have a non-empty file? then # Found it! # Look for the current $FSMOUNT value in the file # using the check_exceptions function defined above. check_exceptions RC=$? # Get the return code from the function if [ $RC -eq 0 ] # Found Exceeded in Exceptions File!! then echo “$FSDEVICE mount on $FSMOUNT is ${FSVALUE}%” \ >> $OUTFILE elif [ $RC -eq 1 ] # Not found in exceptions, use defaults then if [ $FSVALUE -gt $FSMAX ] # Use Script Default then echo “$FSDEVICE mount on $FSMOUNT is ${FSVALUE}%” \ >> $OUTFILE fi fi else # No exceptions file use the script default if [ $FSVALUE -gt $FSMAX ] # Use Script Default then echo “$FSDEVICE mount on $FSMOUNT is ${FSVALUE}%” \ >> $OUTFILE

Listing 5.4 fs_mon_AIX_except.ksh shell script. (continued)

File System Monitoring

109

fi fi done < $WORKFILE # Feed the while loop from the bottom... # Display output if anything is exceeded... if [[ -s $OUTFILE ]] then echo “\nFull Filesystem(s) on ${THISHOST}\n” cat $OUTFILE print fi

Listing 5.4 fs_mon_AIX_except.ksh shell script. (continued)

Notice in the script that we never acted on the return code 2. Because the mount point is found to be OK, there is nothing to do except to check the next mount point. The /usr/local/bin/exceptions file will look something like the script shown in Listing 5.5.

# # # # # # # # # # #

FILE: “exceptions” This file is used to override the $FSMAX value in the filesystem monitoring script fs_mon_excep.ksh. The syntax to override is a /mount-point and a NEW_MAX%: EXAMPLE: /opt 97 OR /usr 96% All lines beginning with a # are ignored as well as the % sign, if you want to use one...

/opt 96% /usr 97 / 50%

Listing 5.5 Example exceptions file.

When we execute the fs_mon_AIX_excep.ksh script, with the exception file entries from Listing 5.5, the output looks like the following on yogi (see Listing 5.6).

Free & Share & Open

110

Chapter 5

Full Filesystem(s) on yogi /dev/hd4 mount on / is 51% /dev/hd10opt mount on /opt is 97%

Listing 5.6 Full filesystem on yogi script in action.

Notice that we added a limit for the root filesystem, /, and set it to 50 percent, and also that this root entry is not at the top of the list in the exceptions file so we have solved the grep problem. You should be able to follow the logic through the preceding code to see that we met all of the goals we set out to accomplish in this section. There are plenty of comments to help you understand each step. Are we finished? Not by a long shot! What about monitoring large filesystems? Using the percentage of filesystem space used is excellent for regular filesystems, but if you have a 10GB filesystem and it is at 90 percent you still have 1GB of free space. Even at 99 percent you have 100MB of space left. For large filesystems we need another monitoring method.

Using the MB of Free Space Method
Sometimes a percentage is just not accurate enough to get the detailed notification that is desired. For these instances, and in the case of large filesystems, we can use awk on the df -k command output to extract the KB of free space field and compare this to a threshold trigger value, specified in either KB or MB. We are going to modify both of the scripts we have already written to use the KB of free space field. Remember our previous df -k command output:
Filesystem 1024-blocks /dev/hd4 32768 /dev/hd2 1212416 /dev/hd9var 53248 /dev/hd3 106496 /dev/hd1 4096 /proc /dev/hd10opt 638976 /dev/scripts_lv 102400 /dev/cd0 656756 Free %Used Iused %Iused Mounted on 16376 51% 1663 11% / 57592 96% 36386 13% /usr 30824 43% 540 5% /var 99932 7% 135 1% /tmp 3916 5% 25 3% /home - /proc 24456 97% 15457 10% /opt 95264 7% 435 2% /scripts 0 100% 328378 100% /cdrom

Instead of the fourth field of the percentage used, we now want to extract the third field with the 1024-blocks, or KB of free space. When someone is working with the script it is best that an easy and familiar measurement is used; the most common is MB

File System Monitoring of free space. To accomplish this we will need to do a little math, but this is just to have a more familiar measurement to work with. As before, we are going to load the command output into the $WORKFILE, but this time we extract columns $1, $3, and $7. df -k | tail +2 | egrep -v ‘/dev/cd[0-9]|/proc’ \ | awk ‘{print $1, $3, $7}’ > $WORKFILE

111

We also need a new threshold variable to use for this method. The MIN_MB_FREE variable sounds good. But what is an appropriate value to set the threshold? In this example we are going to use 50MB. It could be any value, though.
MIN_MB_FREE=”50MB”

Notice that we added MB to the value. We will remove this later, but it is a good idea to add the measurement type just so that the ones who follow will know that the threshold is in MB. Remember that the system is reporting in KB, so we have to multiply our 50MB times 1024 to get the actual value that is equivalent to the systemreported measurement. We also want to strip out the MB letters and typeset the MIN_MB_FREE variable to be an integer. In the compound statement that follows, we take care of everything except typesetting the variable:
(( MIN_MB_FREE = $(echo $MIN_MB_FREE | sed s/MB//g) * 1024 ))

The order of execution for this compound command is as follows: First, the innermost $()command substitution is executed, which replaces the letters MB, if they exist, with null characters. Next is the evaluation of the math equation and assignment of the result to the MIN_MB_FREE variable. Equating MIN_MB_FREE may seem a little confusing, but remember that the system is reporting in KB so we need to get to the same power of 2 to also report in 1024-byte blocks. Other than these small changes, the script is the same as the original, as shown in Listing 5.7.

#!/usr/bin/ksh # # SCRIPT: fs_mon_AIX_MBFREE.ksh # AUTHOR: Randy Michael # DATE: 08-22-2001 # REV: 1.5.P # PURPOSE: This script is used to monitor for full filesystems, # which is defined as “exceeding” the FSMAX value. # A message is displayed for all “full” filesystems. # # REV LIST: # Randy Michael - 08-27-2001

Listing 5.7 fs_mon_AIX_MBFREE.ksh shell script. (continues)

Free & Share & Open

112

Chapter 5

# Changed the code to use MB of free space instead of # the %Used method. # # set -n # Uncomment to check syntax without any execution # set -x # Uncomment to debug this script # ##### DEFINE FILES AND VARIABLES HERE #### MIN_MB_FREE=”50MB” # Min. MB of Free FS Space

WORKFILE=”/tmp/df.work” # Holds filesystem data >$WORKFILE # Initialize to empty OUTFILE=”/tmp/df.outfile” # Output display file >$OUTFILE # Initialize to empty THISHOST=`hostname` # Hostname of this machine ######## START OF MAIN ############# # Get the data of interest by stripping out /dev/cd#, # /proc rows and keeping columns 1, 4 and 7 df -k | tail +2 | egrep -v ‘/dev/cd[0-9] | /proc’ \ | awk ‘{print $1, $3, $7}’ > $WORKFILE # Format Variables (( MIN_MB_FREE = $(echo $MIN_MB_FREE | sed s/MB//g) * 1024 )) # Loop through each line of the file and compare column 2 while read FSDEVICE FSMB_FREE FSMOUNT do FSMB_FREE=$(echo $FSMB_FREE | sed s/MB//g) # Remove the “MB” if (( FSMB_FREE < MIN_MB_FREE )) then (( FS_FREE_OUT = FSMB_FREE / 1000 )) echo “$FSDEVICE mounted on $FSMOUNT only has ${FS_FREE_OUT}MB Free” >> $OUTFILE fi done < $WORKFILE # Feed the while loop from the bottom!!

if [[ -s $OUTFILE ]] then echo “\nFull Filesystem(s) on $THISHOST\n” cat $OUTFILE print fi

Listing 5.7 fs_mon_AIX_MBFREE.ksh shell script. (continued)

File System Monitoring

113

Full Filesystem(s) on yogi /dev/hd4 mounted on / only has 16MB Free /dev/hd9var mounted on /var only has 30MB Free /dev/hd1 mounted on /home only has 3MB Free /dev/hd10opt mounted on /opt only has 24MB Free

Listing 5.8 Shell script in action.

This output in Listing 5.8 is padded by less than 1 MB due to the fact that we divided the KB free column by 1000 for the output, measured in MB. If the exact KB is needed, then the division by 1000 can be omitted. What about giving this script exception capability to raise or lower the threshold, as we did for the percentage technique? We already have the percentage script with the check_exception function so that we can modify this script and function to use the same technique of parsing through the $EXCEPTIONS file.

Using MB of Free Space with Exceptions
To add exception capability to the fs_mon_MBFREE.ksh shell script, we will again need a function to perform the search of the $EXCEPTIONS file, if it exists. This time we will add some extras. We may have the characters MB in our data, so we need to allow for this. We also need to test for null characters, or no data, and remove all blank lines in the exception file. The easiest way to use the function is to supply an appropriate return code back to the calling script. We will set the function up to return 1, one, if the mount point is found to be out of limits in the $DATA_EXCEPTIONS file. It will return 2 if the /mount_point is in the exceptions data file but is not out of limits. The function will return 3 if the mount point is not found in the exceptions data file. This will allow us to call the function to check the exception file, and based on the return code, we make a decision in the main body of the script. We already have experience modifying the script to add exception capability, so this should be a breeze, right? When we finish, the exception modification will be intuitively obvious. Because we are going to parse through the exceptions file, we need to run a sanity check to see if someone made an incorrect entry and placed a colon, :, in the file intending to override the limit on an NFS mounted filesystem. This error should never occur, but because a tester I know did so, I now check and correct the error, if possible. We just cut out the second field using the colon, :, as a delimiter. Listing 5.9 shows the modified check_exceptions function. Check out the highlighted parts in particular.

Free & Share & Open

114

Chapter 5

function check_exceptions { # set -x # Uncomment to debug this function while read FSNAME FSLIMIT do # Do an NFS sanity check echo $FSNAME | grep “:” >/dev/null \ && FSNAME=$(echo $FSNAME | cut -d “:” -f2) # Make sure we do not have a null value if [[ ! -z “$FSLIMIT” && “$FSLIMIT” != ‘’ ]] then (( FSLIMIT = $(echo $FSLIMIT | sed s/MB//g) * 1024 )) if [[ $FSNAME = $FSMOUNT ]] then # Get rid of the “MB” if it exists FSLIMIT=$(echo $FSLIMIT | sed s/MB//g) if (( FSMB_FREE < FSLIMIT )) then return 1 # Found out of limit else return 2 # Found OK fi fi fi done < $DATA_EXCEPTIONS # Feed the loop from the bottom!!! return 3 # Not found in $EXCEPTIONS file }

Listing 5.9 New check_exceptions function.

A few things to notice in this function are the NFS and null value sanity checks as well as the way that we feed the while loop from the bottom, after the done statement. First, the sanity checks are very important to guard against incorrect NFS entries and blank lines, or null data, in the exceptions file. For the NFS colon check we use the double ampersands, &&, as opposed to if...then... statement. It works the same but is cleaner in this type of test. The other point is the null value check. We check for both a zero-length variable and null data. The double ampersands, &&, are called a logical AND function, and the double pipes, ||, are a logical OR function. In a logical AND, &&, all of the command statements must be true for the return code of the entire statement to be 0, zero. In a logical OR, ||, at least one statement must be true for the return code to be 0, zero. When a logical OR receives the first true statement in the test list it will immediately exit the test, or command statement, with a return code of 0, zero.

File System Monitoring
Both are good to use, but some people find it hard to follow. Next we test for an empty/null variable. if [[ ! -z “$FSLIMIT” && “$FSLIMIT” != ‘’ ]]

115

Note that in the null sanity check there are double quotes around both of the $FSLIMIT variables, “$FSLIMIT”. These are required! If you omit the double quotes and the variable is actually null, then the test will fail and a system error message is generated and displayed on the terminal. It never hurts to add double quotes around a variable, and sometimes it is required. For the while loop we go back to our favorite loop structure. Feeding the while loop from the bottom, after done, is the fastest way to loop through a file line by line. With the sanity checks complete, we just compare some numbers and give back a return code to the calling shell script. Please pay attention to the boldface code in Listing 5.10.
#!/usr/bin/ksh # # SCRIPT: fs_mon_AIX_MB_FREE_excep.ksh # AUTHOR: Randy Michael # DATE: 08-22-2001 # REV: 2.1.P # PURPOSE: This script is used to monitor for full filesystems, # which is defined as “exceeding” the FSMAX value. # A message is displayed for all “full” filesystems. # # PLATFORM: AIX # # REV LIST: # Randy Michael - 08-27-2001 # Changed the code to use MB of free space instead of # the %Used method. # # Randy Michael - 08-27-2001 # Added code to allow you to override the set script default # for MIN_MB_FREE of FS Space # # set -n # Uncomment to check syntax without any execution # set -x # Uncomment to debug this script # ##### DEFINE FILES AND VARIABLES HERE #### MIN_MB_FREE=”50MB” # Min. MB of Free FS Space

WORKFILE=”/tmp/df.work” # Holds filesystem data >$WORKFILE # Initialize to empty OUTFILE=”/tmp/df.outfile” # Output display file

Listing 5.10 fs_mon_AIX_MB_FREE_excep.ksh shell script. (continues)

Free & Share & Open

116

Chapter 5

>$OUTFILE # Initialize to empty EXCEPTIONS=”/usr/local/bin/exceptions” # Override data file DATA_EXCEPTIONS=”/tmp/dfdata.out” # Exceptions file w/o # rows THISHOST=`hostname` # Hostname of this machine ####### DEFINE FUNCTIONS HERE ######## function check_exceptions { # set -x # Uncomment to debug this function while read FSNAME FSLIMIT do # Do an NFS sanity check echo $FSNAME | grep “:” >/dev/null \ && FSNAME=$(echo $FSNAME | cut -d “:” -f2) if [[ ! -z “$FSLIMIT” && “$FSLIMIT” != ‘’ ]] # Check for empty/null then (( FSLIMIT = $(echo $FSLIMIT | sed s/MB//g) * 1024 )) if [[ $FSNAME = $FSMOUNT ]] then # Get rid of the “MB” if it exists FSLIMIT=$(echo $FSLIMIT | sed s/MB//g) if (( FSMB_FREE < FSLIMIT )) # Numerical Test then return 1 # Found out of limit else return 2 # Found OK fi fi fi done < $DATA_EXCEPTIONS # Feed the loop from the bottom!!! return 3 # Not found in $EXCEPTIONS file } ######## START OF MAIN ############# # Load the $EXCEPTIONS file if it exists if [[ -s $EXCEPTIONS ]] then # Ignore all lines beginning with a pound sign, # # and omit all blank lines cat $EXCEPTIONS | grep -v “^#” | sed /^$/d > $DATA_EXCEPTIONS fi # Get the data of interest by stripping out /dev/cd#,

Listing 5.10 fs_mon_AIX_MB_FREE_excep.ksh shell script. (continued)

File System Monitoring

117

# /proc rows and keeping columns 1, 4 and 7 df -k | tail +2 | egrep -v ‘/dev/cd[0-9] | /proc’ \ | awk ‘{print $1, $3, $7}’ > $WORKFILE # Format Variables for the proper MB value (( MIN_MB_FREE = $(echo $MIN_MB_FREE | sed s/MB//g) * 1024 )) # Loop through each line of the file and compare column 2 while read FSDEVICE FSMB_FREE FSMOUNT do if [[ -s $EXCEPTIONS ]] then check_exceptions RC=”$?” # Check the Return Code! if (( RC == 1 )) # Found out of exceptions limit then (( FS_FREE_OUT = $FSMB_FREE / 1000 )) echo “$FSDEVICE mounted on $FSMOUNT only has\ ${FS_FREE_OUT}MB Free” \ >> $OUTFILE elif (( RC == 2 )) # Found in exceptions to be OK then # Just a sanity check - We really do nothing here... # The colon, :, is a NO-OP operator in KSH : # No-Op - Do Nothing! elif (( RC == 3 )) # Not found in the exceptions file then FSMB_FREE=$(echo $FSMB_FREE | sed s/MB//g) # Remove the “MB” if (( FSMB_FREE < MIN_MB_FREE )) then (( FS_FREE_OUT = FSMB_FREE / 1000 )) echo “$FSDEVICE mounted on $FSMOUNT only has\ ${FS_FREE_OUT}MB Free” >> $OUTFILE fi fi else # No Exceptions file use the script default FSMB_FREE=$(echo $FSMB_FREE | sed s/MB//g) # Remove the “MB” if (( FSMB_FREE < MIN_MB_FREE )) then (( FS_FREE_OUT = FSMB_FREE / 1000 )) echo “$FSDEVICE mounted on $FSMOUNT only has\ ${FS_FREE_OUT}MB Free” >> $OUTFILE fi fi done < $WORKFILE

Listing 5.10 fs_mon_AIX_MB_FREE_excep.ksh shell script. (continues)

Free & Share & Open

118

Chapter 5

if [[ -s $OUTFILE ]] then echo “\nFull Filesystem(s) on $THISHOST\n” cat $OUTFILE print fi

Listing 5.10 fs_mon_AIX_MB_FREE_excep.ksh shell script. (continued)

The script in Listing 5.10 is good, and we have covered all of the bases, right? If you want to stop here, you will be left with an incomplete picture of what we can accomplish. There are several more things to consider, and, of course, there are many more ways to do any of these tasks, and no one is correct. Let’s consider mixing the filesystem percentage used and the MB of free filesystem space techniques. With a mechanism to auto-detect the way we select the usage, the filesystem monitoring script could be a much more robust tool—and a must-have tool where you have a mix of regular and large filesystems to monitor.

Percentage Used—MB Free and Large Filesystems
Now we’re talking! Even if most of your filesystems are large file enabled or are just huge in size, the small ones will still kill you in the end. For a combination of small and large filesystems, we need a mix of both the percent used and MB of free space techniques. For this combination to work, we need a way to auto-detect the correct usage, which we still need to define. There are different combinations of these auto-detect techniques that can make the monitoring work differently. For the large filesystems we want to use the MB of free space, and for regular filesystems we use the percentage method. We need to define a trigger that allows for this free space versus percentage monitoring transformation. The trigger value will vary by environment, but this example uses 1GB as the transition point from percentage used to MB of free space. Of course, the value should be more like 4–6GB, but we need an example. We also need to consider how the $EXCEPTIONS file is going to look. Options for the exceptions file are a combined file or two separate files, one for percentage used and one for MB free. The obvious choice is one combined file. What are combined entries to look like? How are we going to handle the wrong entry type? The entries need to conform to the specific test type the script is looking for. The best way to handle this is to require that either a % or MB be added as a suffix to each new entry in the exceptions file. With the MB or

File System Monitoring
% suffix we could override not only the triggering level, but also the testing method! If an entry has only a number without the suffix, then this exceptions file entry will be ignored and the shell script’s default values will be used. This suffix method is the most flexible, but it, too, is prone to mistakes in the exceptions file. For the mistakes, we need to test the entries in the exceptions to see that they conform to the standard that we have decided on. The easiest way to create this new, more robust script is to take large portions of the previous scripts and convert them into functions. We can simply insert the word function followed by a function name and enclose the code within curly braces—for example, function test_function { function_code }. Or if you prefer the C-type function method, we can use this example, test_function () { function_code }. The only difference between the two function methods is one uses the word function to define the function while the other just adds a set of parentheses after the function’s name. When we use functions, it is easy to set up a logical framework from which to call the functions. It is always easiest to set up the framework first and then fill in the middle. The logic code for this script will look like Listing 5.11. load_File_System_data > $WORKFILE if EXCEPTIONS_FILE exists and is > 0 size then load_EXCEPTIONS_FILE_data fi while read $WORKFILE, which has the filesystem data do if EXCEPTIONS data was loaded then check_exceptions_file RC=Get Return code back from function case $RC in 1) Found exceeded by % method 2) Found out-of-limit by MB Free method 3) Found OK in exceptions file by a testing method 4) Not found in exceptions file esac else # No exceptions file Use script defaults to compare fi done if we have anything out of limits then display_output fi

119

Listing 5.11 Logic code for a large and small filesystem freespace script.

Free & Share & Open

120

Chapter 5
This is very straightforward and easy to do with functions. From this logical description we already have the main body of the script written. Now we just need to modify the check_exceptions function to handle both types of data and create the load_FS_data, load_EXCEPTIONS_data, and display_output functions. For this script we are also going to do things a little differently because this is a learning process. As we all know, there are many ways to accomplish the same task in Unix; shell scripting is a prime example. To make our scripts a little easier to read at a glance, we are going to change how we do numeric test comparisons. We currently use the standard bracketed test functions with the numeric operators, -lt, -le, -eq, -ne, -ge, and -gt: if [ $VAR1 -gt $VAR2 ]

We are now going to use the bracketed tests for character strings only and do all of our numerical comparisons with the double parentheses method: if (( VAR1 > VAR2 ))

The operators for this method are . When we make this small change, it makes the script much easier to follow because we know immediately that we are dealing with either numeric data or a character string without knowing much at all about the data being tested. Notice that we did not reference the variables with a $ (dollar sign) for the numeric tests. The $ omission is not the only difference, but it is the most obvious. The $ is omitted because it is implied that anything that is not numeric is a variable. Other things to look for in this script are compound tests, math and math within tests, the use of curly braces with variables, ${VAR1}MB, a no-op using a : (colon), data validation, error checking, and error notification. These variables are a lot to look for, but you can learn much from studying the script shown in Listing 5.12. Just remember that all functions must be defined before they can be used! Failure to define functions is the most common mistake when working with them. The second most common mistake has to do with scope. Scope deals with where a variable and its value are known to other scripts and functions. Top level down is the best way to describe where scope lies. The basic rules say that all of a shell script’s variables are known to the internal, lower-level, functions, but none of the function’s variables are known to any higher-calling script or function, thus the top level down definition. We will cover a method called a co-process of dealing with scope in a later chapter. So, in this script the check_exceptions function will use the global script’s variables, which are known to all of the functions, and the function will, in turn, reply with a return code, as we defined in the logic flow of Listing 5.11. Scope is a very important concept, as is the placement of the function in the script. The comments in this script are extensive, so please study the code and pay particular attention to the boldface text.

N OT E Remember: You have to define a function before you can use it.

File System Monitoring

121

#!/usr/bin/ksh # # SCRIPT: fs_mon_AIX_PC_MBFREE_excep.ksh # AUTHOR: Randy Michael # DATE: 08-22-2001 # REV: 4.3.P # PURPOSE: This script is used to monitor for full filesystems, # which is defined as “exceeding” the MAX_PERCENT value. # A message is displayed for all “full” filesystems. # # PLATFORM: AIX # # REV LIST: # Randy Michael - 08-27-2001 # Changed the code to use MB of free space instead of # the %Used method. # # Randy Michael - 08-27-2001 # Added code to allow you to override the set script default # for MIN_MB_FREE of FS Space # # Randy Michael - 08-28-2001 # Changed the code to handle both %Used and MB of Free Space. # It does an “auto-detection” but has override capability # of both the trigger level and the monitoring method using # the exceptions file pointed to by the $EXCEPTIONS variable # # set -n # Uncomment to check syntax without any execution # set -x # Uncomment to debug this script # ##### DEFINE FILES AND VARIABLES HERE #### MIN_MB_FREE=”100MB” MAX_PERCENT=”85%” FSTRIGGER=”1000MB” # Min. MB of Free FS Space # Max. FS percentage value # Trigger to switch from % Used to MB Free

WORKFILE=”/tmp/df.work” # Holds filesystem data >$WORKFILE # Initialize to empty OUTFILE=”/tmp/df.outfile” # Output display file >$OUTFILE # Initialize to empty EXCEPTIONS=”/usr/local/bin/exceptions” # Override data file DATA_EXCEPTIONS=”/tmp/dfdata.out” # Exceptions file w/o # rows EXCEPT_FILE=”N” # Assume no $EXCEPTIONS FILE

Listing 5.12 fs_mon_AIX_PC_MBFREE_excep.ksh shell script. (continues)

Free & Share & Open

122

Chapter 5

THISHOST=`hostname`

# Hostname of this machine

###### FORMAT VARIABLES HERE ###### # Both of these variables need to be multiplied by 1024 blocks (( MIN_MB_FREE = $(echo $MIN_MB_FREE | sed s/MB//g) * 1024 )) (( FSTRIGGER = $(echo $FSTRIGGER | sed s/MB//g) * 1024 )) ####### DEFINE FUNCTIONS HERE ######## function check_exceptions { # set -x # Uncomment to debug this function while read FSNAME FSLIMIT do IN_FILE=”N” # If found in file, which test type to use? # Do an NFS sanity check and get rid of any “:”. # If this is found it is actually an error entry # but we will try to resolve it. It will # work only if it is an NFS cross mount to the same # mount point on both machines. echo $FSNAME | grep ‘:’ >/dev/null \ && FSNAME=$(echo $FSNAME | cut -d ‘:’ -f2) # Check for empty and null variable if [[ ! -z “$FSLIMIT” && “$FSLIMIT” != ‘’ ]] then if [[ $FSNAME = $FSMOUNT ]] # Found it! then # Check for “MB” Characters...Set IN_FILE=MB echo $FSLIMIT | grep MB >/dev/null && IN_FILE=”MB” \ && (( FSLIMIT = $(echo $FSLIMIT \ | sed s/MB//g) * 1024 )) # check for “%” Character...Set IN_FILE=PC, for % echo $FSLIMIT | grep “%” >/dev/null && IN_FILE=”PC” \ && FSLIMIT=$(echo $FSLIMIT | sed s/\%//g) case $IN_FILE in MB) # Use Megabytes of free space to test # Up-case the characters, if they exist FSLIMIT=$(echo $FSLIMIT | tr ‘[a-z]’ ‘[A-Z]’) # Get rid of the “MB” if it exists FSLIMIT=$(echo $FSLIMIT | sed s/MB//g) # Test for blank and null values if [[ ! -z $FSLIMIT && $FSLIMIT != ‘’ ]]

Listing 5.12 fs_mon_AIX_PC_MBFREE_excep.ksh shell script. (continued)

File System Monitoring

123

then # Test for a valid filesystem “MB” limit if (( FSLIMIT >= 0 && FSLIMIT < FSSIZE )) then # Check the limit if (( FSMB_FREE < FSLIMIT )) then return 1 # Found out of limit # using MB Free method else return 3 # Found OK fi else echo “\nERROR: Invalid filesystem MAX for\ $FSMOUNT - $FSLIMIT” echo “ Exceptions file value must be\ less than or” echo “ equal to the size of the filesystem\ measured” echo “ in 1024 bytes\n” fi else echo “\nERROR: Null value specified in exceptions\ file” echo “ for the $FSMOUNT mount point.\n” fi ;; PC) # Use the Percent used method to test # Strip out the % sign if it exists PC_USED=$(echo $PC_USED | sed s/\%//g) # Test for blank and null values if [[ ! -z $FSLIMIT && $FSLIMIT != ‘’ ]] then # Test for a valid percentage, i.e. 0-100 if (( FSLIMIT >= 0 && FSLIMIT FSLIMIT )) then return 2 # Found exceeded by % Used method else return 3 # Found OK fi else echo “\nERROR: Invalid percentage for\ $FSMOUNT - $FSLIMIT” echo “ Exceptions file values must be” echo “ between 0 and 100%\n” fi

Listing 5.12 fs_mon_AIX_PC_MBFREE_excep.ksh shell script. (continues)

Free & Share & Open

124

Chapter 5

N)

else echo “\nERROR: Null value specified in exceptions” echo “ file for the $FSMOUNT mount point.\n” fi ;; # Test type not specified in exception file, use default # Inform the user of the exceptions file error... echo “\nERROR: Missing testing type in exceptions file” echo “ for the $FSMOUNT mount point. A \”%\” or” echo “ \”MB\” must be a suffix to the numerical” echo “ entry. Using script default values...\n” # Method Not Specified - Use Script Defaults if (( FSSIZE >= FSTRIGGER )) then # This is a “large” filesystem if (( FSMB_FREE < MIN_MB_FREE )) then return 1 # Found out of limit using MB Free else return 3 # Found OK fi else # This is a standard filesystem PC_USED=$(echo $PC_USED | sed s/\%//g) #Remove the % FSLIMIT=$(echo $FSLIMIT | sed s/\%//g) #Remove the % if (( PC_USED > FSLIMIT )) then return 2 # Found exceeded by % Used method else return 3 # Found OK fi fi ;;

esac fi fi done < $DATA_EXCEPTIONS # Feed the loop from the bottom!!! return 4 # Not found in $EXCEPTIONS file } #################################### function display_output { if [[ -s $OUTFILE ]] then

Listing 5.12 fs_mon_AIX_PC_MBFREE_excep.ksh shell script. (continued)

File System Monitoring

125

echo “\nFull Filesystem(s) on $THISHOST\n” cat $OUTFILE print fi } #################################### function load_EXCEPTIONS_data { # Ignore any line that begins with a pound sign, # # and omit all blank lines cat $EXCEPTIONS | } grep -v “^#” | sed /^$/d > $DATA_EXCEPTIONS

#################################### function load_FS_data { df -k | tail +2 | egrep -v ‘/dev/cd[0-9]|/proc’ \ | awk ‘{print $1, $2, $3, $4, $7}’ > $WORKFILE } #################################### ######### START OF MAIN ############ #################################### load_FS_data # Do we have a nonzero size $EXCEPTIONS file? if [[ -s $EXCEPTIONS ]] then # Found a nonempty $EXCEPTIONS file load_EXCEPTIONS_data EXCEP_FILE=”Y” fi while read FSDEVICE FSSIZE FSMB_FREE PC_USED FSMOUNT do if [[ $EXCEP_FILE = “Y” ]] then check_exceptions CE_RC=”$?” # Check Exceptions Return Code (CE_RC) case $CE_RC in

Listing 5.12 fs_mon_AIX_PC_MBFREE_excep.ksh shell script. (continues)

Free & Share & Open

126

Chapter 5

1) # Found exceeded in exceptions file by MB Method (( FS_FREE_OUT = FSMB_FREE / 1000 )) echo “$FSDEVICE mounted on $FSMOUNT has ${FS_FREE_OUT}MB\ Free” \ >> $OUTFILE ;; 2) # Found exceeded in exceptions file by %Used method echo “$FSDEVICE mount on $FSMOUNT is ${PC_USED}%” \ >> $OUTFILE ;; 3) # Found OK in exceptions file : # NO-OP Do Nothing ;; 4) # Not found in exceptions file - Use Script Default Triggers if (( FSSIZE >= FSTRIGGER )) then # This is a “large” filesystem # Remove the “MB”, if it exists FSMB_FREE=$(echo $FSMB_FREE | sed s/MB//g) typeset -i FSMB_FREE if (( FSMB_FREE < MIN_MB_FREE )) then (( FS_FREE_OUT = FSMB_FREE / 1000 )) echo “$FSDEVICE mounted on $FSMOUNT has\ ${FS_FREE_OUT}MB Free” >> $OUTFILE fi else # This is a standard filesystem PC_USED=$(echo $PC_USED | sed s/\%//g) MAX_PERCENT=$(echo $MAX_PERCENT | sed s/\%//g) if (( PC_USED > MAX_PERCENT )) then echo “$FSDEVICE mount on $FSMOUNT is ${PC_USED}%” \ >> $OUTFILE fi fi ;; esac else # NO $EXECPTIONS FILE USE DEFAULT TRIGGER VALUES if (( FSSIZE >= FSTRIGGER )) then # This is a “large” filesystem - Use MB Free Method FSMB_FREE=$(echo $FSMB_FREE | sed s/MB//g) # Remove the “MB” if (( FSMB_FREE < MIN_MB_FREE )) then (( FS_FREE_OUT = FSMB_FREE / 1000 )) echo “$FSDEVICE mounted on $FSMOUNT has\

Listing 5.12 fs_mon_AIX_PC_MBFREE_excep.ksh shell script. (continued)

File System Monitoring

127

${FS_FREE_OUT}MB Free” >> $OUTFILE fi else # This is a standard filesystem - Use % Used Method PC_USED=$(echo $PC_USED | sed s/\%//g) MAX_PERCENT=$(echo $MAX_PERCENT | sed s/\%//g) if (( PC_USED > MAX_PERCENT )) then echo “$FSDEVICE mount on $FSMOUNT is ${PC_USED}%” \ >> $OUTFILE fi fi fi done < $WORKFILE # Feed the while loop from the bottom!!! display_output # End of Script

Listing 5.12 fs_mon_AIX_PC_MBFREE_excep.ksh shell script. (continued)

In the script shown in Listing 5.12, we made tests to confirm the data’s integrity and for mistakes in the exceptions file (of course, we can go only so far with mistakes!). The reason is that we made the exceptions file more complicated to use. Two of my testers consistently had reverse logic on the MB free override option of the script by thinking greater than instead of less than. From this confusion, a new exceptions file was created that explained what the script is looking for and gave example entries. Of course, all of these lines begin with a pound sign, #, so they are ignored when data is loaded into the $DATA_EXCEPTIONS file. Listing 5.13 shows the exceptions file that worked best with the testers.
# # # # # # # # # # # # # FILE: “exceptions” This file is used to override both the default trigger value in the filesystem monitoring script fs_mon_excep.ksh, but also allows overriding the monitoring technique used, i.e. Max %Used and minimum MB of filesystem space. The syntax to override is a /mount-point and a trigger value. EXAMPLES: /usr 96% OR # Flag anything ABOVE 96%

Listing 5.13 Example exceptions file. (continues)

Free & Share & Open

128

Chapter 5

# /usr 50MB # Flag anything BELOW 50 Megabytes # # All lines beginning with a # are ignored. # # NOTE: All Entries MUST have either “MB” or # “%” as a suffix!!! Or else the script # defaults are used. NO SPACES PLEASE! # /opt 95% / 50% /usr 70MB

Listing 5.13 Example exceptions file. (continued)

The requirement for either % or MB does help keep the entry mistakes down. In case mistakes are made, the error notifications seemed to get these cleared up very quickly—usually after an initial run. You can find customized shell scripts for each of the operating systems (AIX, HP-UX, Linux, and SunOS) on this book’s Web site. Are we finished with filesystem monitoring? No way! What about the other three operating systems that we want to monitor? We need to be able to execute this script on AIX, Linux, HP-UX, and Solaris without the need to change the script on each platform.

Running on AIX, Linux, HP-UX, and Solaris
Can we run the filesystem scripts on various Unix flavors? You bet! Running our filesystem monitoring script is very easy because we used functions for most of the script. We are going to use the same script, but instead of hard-coding the loading of the filesystem data, we need to use variables to point to the correct OS syntax and columns of interest. Now we need a new function that will determine which flavor of Unix we are running. Based on the OS, we set up the command syntax and command output columns of interest that we want to extract and load the filesystem data for this particular OS. For OS determination we just use the uname command. uname, and the get_OS_info function, will return the resident operating system, as shown in Table 5.1.
Table 5.1 uname Command and Function Results COMMAND RESULT Linux AIX HP-UX SunOS FUNCTION RESULT LINUX AIX HP-UX SUNOS

OPERATING SYSTEM Linux AIX HP-UX Solaris

File System Monitoring
For the function’s output we want to use all UPPERCASE characters, which makes testing much easier. In the following function please notice we use the typeset function to ensure that the result is in all uppercase characters. function get_OS_info { # For a few commands it is necessary to know the OS to # execute the proper command syntax. This will always # return the Operating System in UPPERCASE characters typeset -u OS OS=`uname` print $OS } # Use the UPPERCASE values for the OS variable # Grab the Operating system, i.e. AIX, HP-UX # Send back the UPPERCASE value

129

To use the get_OS_info function we can assign it to a variable using command substitution, use the function directly in a command statement, or redirect the output to a file. For this script modification we are going to use the get_OS_info function directly in a case statement. Now we need four different load_FS_data functions, one for each of the four operating systems, and that is all of the modification that is needed. Each of the load_FS_data functions will be unique in command syntax and the column fields to extract from the df statement output, as well as the devices to exclude from testing. Because we wrote this script using functions, we will replace the original load_FS_data script, at the Beginning of Main, with a case statement that utilizes the get_OS_info function. The case statement will execute the appropriate load_FS_data function.

case $(get_OS_info) in AIX) # Load filesystem data for AIX load_AIX_FS_data ;; HP-UX) # Load filesystem data for HP-UX load_HP_UX_FS_data ;; LINUX) # Load filesystem data for Linux load_LINUX_FS_data ;; SUNOS) # Load filesystem data for Solaris load_Solaris_FS_data ;; *) # Unsupported in script echo “\nUnsupported Operating System...EXITING\n” exit 1 esac

Listing 5.14 Operating system test.

Free & Share & Open

130

Chapter 5
Listing 5.14 shows simple enough replacement code. In this case statement we either execute one of the functions or exit if the OS is not in the list with a return code of 1, one. In these functions we will want to pay attention to the command syntax for each operating system, the columns to extract for the desired data, and the filesystems that we want to ignore, if any. There is an egrep, or extended grep, in each statement that will allow for exclusions to the filesystems that are monitored. A typical example of this is a CD-ROM. Remember that a CD-ROM will always show that it is 100% utilized because it is mounted as read-only and you cannot write to it. Also, some operating systems list mount points that are really not meant to be monitored, such as /proc in AIX 5L.

Command Syntax and Output Varies between Operating Systems
The command syntax and command output varies between Unix operating systems. To get a similar output of the AIX df -k command on other operating systems we sometimes have to change the command syntax. We also extract data from different columns in the output. The command syntax and resulting output for AIX, Linux, HPUX, and SUN/Solaris are listed in the text that follows as well as the columns of interest for each operating system output. Please review Tables 5.2 through 5.9.

Table 5.2

AIX df -k Command Output 1024BLOCKS 32768 1212416 53248 106496 4096 FREE 16376 57592 30824 99932 3916 %USED 51% 96% 43% 7% 5% IUSED 1663 36386 540 135 25 %IUSED 11% 13% 5% 1% % MOUNTED ON / /usr /var /tmp /home /proc 638976 102400 656756 24456 95264 0 97% 7% 100% 15457 435 10% 2% /opt /scripts /cdrom

FILESYSTEM /dev/hd4 /dev/hd2 /dev/hd9var /dev/hd3 /dev/hd1 /proc /dev/hd10opt /dev/scripts_lv /dev/cd0

328378 100%

File System Monitoring
Table 5.3 AIX df Output Columns of Interest DF OUTPUT COLUMNS Column 1 Column 2 Column 3 Column 4 Column 7 COLUMN CONTENTS The filesystem device name, Filesystem The size of the filesystem in 1024 blocks, 1024-blocks The kilobytes of free filesystem space, Free The percentage of used capacity, %Used The mount point of the filesystem, Mounted on

131

Table 5.4 Linux df -k Command Output FILESYSTEM /dev/hda16 /dev/hda5 /dev/hda1 /dev/hda8 /dev/hda9 /dev/hda12 /dev/hda6 /dev/hda10 /dev/hda11 1K-BLOCKS 101089 1011928 54416 202220 202220 124427 1011928 155545 124427 USED 32949 104 2647 13 1619 19 907580 36 29670 AVAILABLE 62921 960420 48960 191767 190161 117984 52944 147479 88333 USE% 34% 0% 5% 0% 1% 0% 94% /usr 0% 25%/var MOUNTED ON / /backup /boot /download /home /tmp — /usr/local

Table 5.5 Linux df Output Columns of Interest DF OUTPUT COLUMNS Column 1 Column 2 Column 4 Column 5 Column 6 COLUMN CONTENTS The filesystem device name, Filesystem The size of the filesystem in 1k-blocks, 1k-blocks The kilobytes of free filesystem space, Available The percentage of used capacity, Use% The mount point of the filesystem, Mounted on

Free & Share & Open

132

Chapter 5
Table 5.6 SUN/Solaris df -k Command Output KBYTES 192423 1015542 0 0 0 96455 554132 47975 554428 1015542 375255 USED 18206 488678 0 0 0 5931 0 1221 296 1 214843 AVAIL 154975 465932 0 0 0 80879 55413 41957 554132 954598 122887 MOUNTED CAPACITY ON 11% 52% 0% 0% 0% 7% 0% 3% 1% 1% 64% /usr /proc /dev/fd /etc/mnttab /var /var/run /opt /tmp /export/home /usr/openwin

FILESYSTEM /dev/dsk/c0d0s0 /dev/dsk/c0d0s6 /proc fd mnttab /dev/dsk/c0d0s3 swap /dev/dsk/c0d0s5 swap /dev/dsk/c0d0s7 /dev/dsk/c0d0s1

Table 5.7

SUN/Solaris df Òk Output Columns of Interest COLUMN CONTENTS The filesystem device name, Filesystem The size of the filesystem in 1k-blocks, kbytes The kilobytes of free filesystem space, avail The percentage of used capacity, capacity The mount point of the filesystem, Mounted on

DF OUTPUT COLUMNS Column 1 Column 2 Column 4 Column 5 Column 6

Table 5.8

HP-UX bdf Command Output KBYTES 151552 47829 1310720 972800 USED 89500 24109 860829 554392 AVAIL 58669 18937 422636 392358 %USED 60% 56% 67% 59% MOUNTED ON / /stand /var /usr

FILESYSTEM /dev/vg00/lvol3 /dev/vg00/lvol1 /dev/vg00/lvol9 /dev/vg00/lvol8

File System Monitoring
Table 5.8 (Continued) FILESYSTEM /dev/vg13/lvol1 /dev/vg00/lvol7 /dev/vg00/lvol13 /dev/vg00/lvol6 /dev/vg00/lvol5 KBYTES 4190208 102400 2039808 720896 409600 USED 1155095 4284 1664073 531295 225464 AVAIL 2850597 92256 352294 177953 176663 %USED 29% 4% 83% 75% 56% MOUNTED ON /u2 /tmp /test2 /opt /home

133

Table 5.9

HP-UX bdf Output Columns of Interest COLUMN CONTENTS The filesystem device name, Filesystem The size of the filesystem in 1k-blocks, kbytes The kilobytes of free filesystem space, avail The percentage of used capacity, %used The mount point of the filesystem, Mounted on

DF OUTPUT COLUMNS Column 1 Column 2 Column 4 Column 5 Column 6

Now that we know how the commands and output vary between operating systems, we can take this into account when creating the shell functions to load the correct filesystem data for each system. Note in each of the following functions that one or more filesystems or devices are set to be ignored, which is specified by the egrep part of the statement.
#################################### function load_AIX_FS_data { df -k | tail +2 | egrep -v ‘/dev/cd[0-9]|/proc’ \ | awk ‘{print $1, $2, $3, $4, $7}’ > $WORKFILE } #################################### function load_HP_UX_FS_data { bdf | tail +2 | egrep -v ‘/mnt/cdrom’ \ | awk ‘{print $1, $2, $4, $5, $6}’ > $WORKFILE

Free & Share & Open

134

Chapter 5
} #################################### function load_LINUX_FS_data { df -k | tail +2 | egrep -v ‘/mnt/cdrom’\ | awk ‘{print $1, $2, $4, $5, $6}’ > $WORKFILE } #################################### function load_Solaris_FS_data { df -k | tail +2 | egrep -v ‘/dev/fd|/etc/mnttab|/proc’\ | awk ‘{print $1, $2, $4, $5, $6}’ > $WORKFILE }

Each Unix system is different, and these functions may need to be modified for your particular environment. The script modification to execute on all of the four operating systems includes entering the functions into the top part of the script, where functions are defined, and to replace the current load_FS_data function with a case statement that utilizes the get_OS_info function. This is an excellent example of how using functions can make life doing modifications much easier. The final script (it is never a final script!) will look like the following code, shown in Listing 5.15. Please scan through the boldface text in detail.

#!/usr/bin/ksh # # SCRIPT: fs_mon_ALL_OS.ksh # AUTHOR: Randy Michael # DATE: 08-22-2001 # REV: 5.1.D # # PURPOSE: This script is used to monitor for full filesystems, # which are defined as “exceeding” the MAX_PERCENT value. # A message is displayed for all “full” filesystems. # # PLATFORM: AIX, Linux, HP-UX and Solaris # # REV LIST: # Randy Michael - 08-27-2001 # Changed the code to use MB of free space instead of # the %Used method. # # Randy Michael - 08-27-2001 # Added code to allow you to override the set script default

Listing 5.15 fs_mon_ALL_OS.ksh shell script.

File System Monitoring

135

# for MIN_MB_FREE of FS Space # # Randy Michael - 08-28-2001 # Changed the code to handle both %Used and MB of Free Space. # It does an “auto-detection” but has override capability # of both the trigger level and the monitoring method using # the exceptions file pointed to by the $EXCEPTIONS variable # # Randy Michael - 08-28-2001 # Added code to allow this script to be executed on # AIX, Linux, HP-UX, and Solaris # # set -n # Uncomment to check syntax without any execution # set -x # Uncomment to debug this script # ##### DEFINE FILES AND VARIABLES HERE #### MIN_MB_FREE=”100MB” MAX_PERCENT=”85%” FSTRIGGER=”1000MB” # Min. MB of Free FS Space # Max. FS percentage value # Trigger to switch from % Used to MB Free

WORKFILE=”/tmp/df.work” # Holds filesystem data >$WORKFILE # Initialize to empty OUTFILE=”/tmp/df.outfile” # Output display file >$OUTFILE # Initialize to empty EXCEPTIONS=”/usr/local/bin/exceptions” # Override data file DATA_EXCEPTIONS=”/tmp/dfdata.out” # Exceptions file w/o # rows EXCEPT_FILE=”N” # Assume no $EXCEPTIONS FILE THISHOST=`hostname` # Hostname of this machine ###### FORMAT VARIABLES HERE ###### # Both of these variables need to be multiplied by 1024 blocks (( MIN_MB_FREE = $(echo $MIN_MB_FREE | sed s/MB//g) * 1024 )) (( FSTRIGGER = $(echo $FSTRIGGER | sed s/MB//g) * 1024 )) ###################################### ####### DEFINE FUNCTIONS HERE ######## ###################################### function get_OS_info { # For a few commands it is necessary to know the OS and its level # to execute the proper command syntax. This will always return # the OS in UPPERCASE typeset -u OS # Use the UPPERCASE values for the OS variable OS=`uname` # Grab the Operating system, i.e. AIX, HP-UX

Listing 5.15 fs_mon_ALL_OS.ksh shell script. (continues)

Free & Share & Open

136

Chapter 5

print $OS }

# Send back the UPPERCASE value

#################################### function check_exceptions { # set -x # Uncomment to debug this function while read FSNAME FSLIMIT do IN_FILE=”N” # Do an NFS sanity check and get rid of any “:”. # If this is found it is actually an error entry # but we will try to resolve it. It will only # work if it is an NFS cross mount to the same # mount point on both machines. echo $FSNAME | grep ‘:’ >/dev/null \ && FSNAME=$(echo $FSNAME | cut -d ‘:’ -f2) # Check for empty and null variable if [[ ! -z $FSLIMIT && $FSLIMIT != ‘’ ]] then if [[ $FSNAME = $FSMOUNT ]] # Found it! then # Check for “MB” Characters...Set IN_FILE=MB echo $FSLIMIT | grep MB >/dev/null && IN_FILE=”MB” \ && (( FSLIMIT = $(echo $FSLIMIT \ | sed s/MB//g) * 1024 )) # check for “%” Character...Set IN_FILE=PC, for % echo $FSLIMIT | grep “%” >/dev/null && IN_FILE=”PC” \ && FSLIMIT=$(echo $FSLIMIT | sed s/\%//g) case $IN_FILE in MB) # Use MB of Free Space Method # Up-case the characters, if they exist FSLIMIT=$(echo $FSLIMIT | tr ‘[a-z]’ ‘[A-Z]’) # Get rid of the “MB” if it exists FSLIMIT=$(echo $FSLIMIT | sed s/MB//g) # Test for blank and null values if [[ ! -z $FSLIMIT && $FSLIMIT != ‘’ ]] then # Test for a valid filesystem “MB” limit if (( FSLIMIT >= 0 && FSLIMIT < FSSIZE )) then if (( FSMB_FREE < FSLIMIT )) then

Listing 5.15 fs_mon_ALL_OS.ksh shell script. (continued)

File System Monitoring

137

return 1 # Found out of limit # using MB Free method else return 3 # Found OK fi else echo “\nERROR: Invalid filesystem MAX for\ $FSMOUNT - $FSLIMIT” echo “ Exceptions file value must be less\ than or” echo “ equal to the size of the filesystem\ measured” echo “ in 1024 bytes\n” fi else echo “\nERROR: Null value specified in exceptions\ file” echo “ for the $FSMOUNT mount point.\n” fi ;; PC) # Use Filesystem %Used Method # Strip out the % sign if it exists PC_USED=$(echo $PC_USED | sed s/\%//g) # Test for blank and null values if [[ ! -z $FSLIMIT && $FSLIMIT != ‘’ ]] then # Test for a valid percentage, i.e. 0-100 if (( FSLIMIT >= 0 && FSLIMIT $FSLIMIT )) then return 2 # Found exceeded by % Used method else return 3 # Found OK fi else echo “\nERROR: Invalid percentage for $FSMOUNT -\ $FSLIMIT” echo “ Exceptions file values must be” echo “ between 0 and 100%\n” fi else echo “\nERROR: Null value specified in exceptions\ file” echo “ for the $FSMOUNT mount point.\n” fi ;; N) # Method Not Specified - Use Script Defaults

Listing 5.15 fs_mon_ALL_OS.ksh shell script. (continues)

Free & Share & Open

138

Chapter 5

if (( FSSIZE >= FSTRIGGER )) then # This is a “large” filesystem if (( FSMB_FREE < MIN_MB_FREE )) then return 1 # Found out of limit # using MB Free method else return 3 # Found OK fi else # This is a standard filesystem PC_USED=$(echo $PC_USED | sed s/\%//g) # Remove % FSLIMIT=$(echo $FSLIMIT | sed s/\%//g) # Remove % if (( PC_USED > FSLIMIT )) then return 2 # Found exceeded by % Used method else return 3 # Found OK fi fi ;; esac fi fi done < $DATA_EXCEPTIONS # Feed the loop from the bottom!!! return 4 # Not found in $EXCEPTIONS file } #################################### function display_output { if [[ -s $OUTFILE ]] then echo “\nFull Filesystem(s) on $THISHOST\n” cat $OUTFILE print fi } #################################### function load_EXCEPTIONS_data { # Ignore any line that begins with a pound sign, # # and omit all blank lines cat $EXCEPTIONS | grep -v “^#” | sed /^$/d > $DATA_EXCEPTIONS

Listing 5.15 fs_mon_ALL_OS.ksh shell script. (continued)

File System Monitoring

139

} #################################### function load_AIX_FS_data { df -k | tail +2 | egrep -v ‘/dev/cd[0-9]|/proc’ \ | awk ‘{print $1, $2, $3, $4, $7}’ > $WORKFILE } #################################### function load_HP_UX_FS_data { bdf | tail +2 | egrep -v ‘/cdrom’ \ | awk ‘{print $1, $2, $4, $5, $6}’ > $WORKFILE } #################################### function load_LINUX_FS_data { df -k | tail +2 | egrep -v ‘/cdrom’\ | awk ‘{print $1, $2, $4, $5, $6}’ > $WORKFILE } #################################### function load_Solaris_FS_data { df -k | tail +2 | egrep -v ‘/dev/fd|/etc/mnttab|/proc’\ | awk ‘{print $1, $2, $4, $5, $6}’ > $WORKFILE } #################################### ######### START OF MAIN ############ #################################### # Query the operating system to find the Unix flavor, then # load the correct filesystem data for the resident OS case $(get_OS_info) in AIX) # Load filesystem data for AIX load_AIX_FS_data

Listing 5.15 fs_mon_ALL_OS.ksh shell script. (continues)

Free & Share & Open

140

Chapter 5

;; HP-UX) # Load filesystem data for HP-UX load_HP_UX_FS_data ;; LINUX) # Load filesystem data for Linux load_LINUX_FS_data ;; SUNOS) # Load filesystem data for Solaris load_Solaris_FS_data ;; *) # Unsupported in script echo “\nUnsupported Operating System for this\ Script...EXITING\n” exit 1 esac # Do we have a nonzero size $EXCEPTIONS file? if [[ -s $EXCEPTIONS ]] then # Found a nonempty $EXCEPTIONS file load_EXCEPTIONS_data EXCEP_FILE=”Y” fi while read FSDEVICE FSSIZE FSMB_FREE PC_USED FSMOUNT do if [[ $EXCEP_FILE = “Y” ]] then check_exceptions CE_RC=”$?” # Check Exceptions Return Code (CE_RC) case $CE_RC in 1) # Found exceeded in exceptions file by MB Method (( FS_FREE_OUT = FSMB_FREE / 1000 )) echo “$FSDEVICE mounted on $FSMOUNT has ${FS_FREE_OUT}MB\ Free” >> $OUTFILE ;; 2) # Found exceeded in exceptions file by %Used method echo “$FSDEVICE mount on $FSMOUNT is ${PC_USED}%” \ >> $OUTFILE ;; 3) # Found OK in exceptions file : # NO-OP Do Nothing. A “:” is a no-op! ;; 4) # Not found in exceptions file - Use Default Triggers if (( FSSIZE >= FSTRIGGER ))

Listing 5.15 fs_mon_ALL_OS.ksh shell script. (continued)

File System Monitoring

141

then # This is a “large” filesystem FSMB_FREE=$(echo $FSMB_FREE | sed s/MB//g) # Remove the\ “MB” if (( FSMB_FREE < MIN_MB_FREE )) then (( FS_FREE_OUT = FSMB_FREE / 1000 )) echo “$FSDEVICE mounted on $FSMOUNT has {FS_FREE_OUT}MB\ Free” >> $OUTFILE fi else # This is a standard filesystem PC_USED=$(echo $PC_USED | sed s/\%//g) MAX_PERCENT=$(echo $MAX_PERCENT | sed s/\%//g) if (( PC_USED > MAX_PERCENT )) then echo “$FSDEVICE mount on $FSMOUNT is ${PC_USED}%” \ >> $OUTFILE fi fi ;; esac else # NO $EXCEPTIONS FILE USE DEFAULT TRIGGER VALUES if (( FSSIZE >= FSTRIGGER )) then # This is a “large” filesystem - Use MB Free Method FSMB_FREE=$(echo $FSMB_FREE | sed s/MB//g) # Remove the “MB” if (( FSMB_FREE < MIN_MB_FREE )) then (( FS_FREE_OUT = FSMB_FREE / 1000 )) echo “$FSDEVICE mounted on $FSMOUNT has ${FS_FREE_OUT}MB Free” \ >> $OUTFILE fi else # This is a standard filesystem - Use % Used Method PC_USED=$(echo $PC_USED | sed s/\%//g) MAX_PERCENT=$(echo $MAX_PERCENT | sed s/\%//g) if (( PC_USED > MAX_PERCENT )) then echo “$FSDEVICE mount on $FSMOUNT is ${PC_USED}%” \ >> $OUTFILE fi fi fi done < $WORKFILE # Feed the while loop from the bottom!!!!! display_output # End of Script

Listing 5.15 fs_mon_ALL_OS.ksh shell script. (continued)

Free & Share & Open

142

Chapter 5
A good study of the script in Listing 5.15 will reveal some nice ways to handle the different situations we encounter while writing shell scripts. As always, it is intuitively obvious! The /usr/local/bin/exceptions file in Listing 5.16 is used on yogi.
# FILE: “exceptions” # # This file is used to override the default # trigger value in the filesystem monitoring script # fs_mon_ALL_OS_excep.ksh, but also allows overriding the # monitoring technique used, i.e. Max %Used and # MINIMUM MB FREE of filesystem space. The syntax to # override is a /mount-point and a “trigger value” with # either “%” or “MB” as a suffix. # # EXAMPLES: # # /usr 96% # OR # /usr 50MB # # All lines beginning with a # are ignored. # # NOTE: All Entries MUST have either “MB” or # “%” as a suffix!!! Or else the script # defaults are used. NO SPACES PLEASE! # /opt 95% / 50% /usr 70MB /home 50MB

Listing 5.16 Sample exceptions file.

Listing 5.16 should work, but it gives an error. If the monitoring script is executed using these exception file entries, it will result in the following output:
ERROR: Invalid filesystem MINIMUM_MB_FREE specified for /home - 50MB -- Current size is 4MB. Exceptions file value must be less than or equal to the size of the filesystem measured Megabytes

Full Filesystem(s) on yogi /dev/hd4 mount on / is 51% /dev/hd2 mounted on /usr has 57MB Free /dev/hd10opt mount on /opt is 97%

File System Monitoring
The problem is with the /home filesystem entry in the $EXCEPTIONS file. The value specified is 50 Megabytes, and the /home filesystem is only 4MB in size. In a case like this the check_exceptions function will display an error message and then use the shell script default values to measure the filesystem and return an appropriate return code to the calling script. So, if a modification is made to the exceptions file, the script needs to be run to check for any errors. The important thing to note is that error checking and data validation should take place before the data is used for measurement. This sequence will also prevent any messages from standard error (stderr) that the system may produce.

143

Other Options to Consider
We can always improve on a script, and the full filesystems script is no exception.

Event Notification
Because monitoring for full filesystems should involve event notification, it is wise to modify the display_output function to send some kind of message, whether by page or email, or otherwise this information needs to be made known so that we can call ourselves proactive. Sending an email to your pager and desktop would be a good start. An entry like the statement that follows might work, but its success depends on the mail server and firewall configurations. echo “Full Filesystem(s) on $THISHOST\n” > $MAILFILE cat $OUTFILE >> $MAILFILE mailx -s “Full Filesystem(s) on $THISHOST” $MAIL_LIST < $MAILFILE

For pager notification, the text message must be very short, but descriptive enough to get the point across.

Automated Execution
If we are to monitor the system, we want the system to tell us when it has a problem. We want event notification, but we also want the event notification to be automated. For filesystem monitoring, a cron table entry is the best way to do this. An interval of about 10–15 minutes 24 × 7 is most common. We have the exceptions capability built in so that if pages become a problem, the exceptions file can be modified to stop the filesystem from being in error, and thus stop the paging. The cron entry that follows will execute the script every 10 minutes, on the 5s, 24 hours a day, 7 days a week.
5,15,25,35,45,55 * * * * /usr/local/bin/fs_mon_ALL_OS.ksh 2>&1

To make this cron entry you can either edit a cron table with crontab -e or use the following command sequence to append an entry to the end of the cron table.

Free & Share & Open

144

Chapter 5 crontab -l > /tmp/cron_hold.out echo ‘5,15,25,35,45,55 * * * * /usr/local/bin/fs_mon_ALL_OS.ksh 2>&1’ \ >> /tmp/cron_hold.out crontab /tmp/cron_hold.out rm /tmp/cron_hold.out

For this to work, the fs_mon_ALL_OS.ksh script must be modified to send notification by some method. Paging, email, SNMP traps, and modem dialing are the preferred methods. You could send this output to the systems console, but who would ever see it?

Modify the egrep Statement
It may be wise to remove the egrep part of the df statement, used for filesystem exclusion, and use another method. As pointed out previously, grepping can be a mistake. Grepping was done here because most of the time we can get a unique character string for a filesystem device to make grep and egrep work without error, but not always. If this is a problem, then creating a list either in a variable assignment in the script or in a file is the best bet. Then the new $IGNORE_LIST list can be searched and an exact match can be made.

Summary
Through this chapter we have changed our thinking about monitoring for full filesystems. The script that we use can be very simple for the average small shop or more complex as we move to larger and larger storage solutions. All filesystems are not created equal in size, and when you get a mix of large and small filesystems on mixed operating systems, we have shown how to handle the mix with ease. In the next chapter we will move into monitoring the paging and/or swap space. If we run out of paging or swap space, the system will start thrashing, and if the problem is chronic, the system may crash. We will look at the different monitoring methods for each operating system.

CHAPTER

6
Monitoring Paging and Swap Space

Every Systems Administrator loves paging and swap space because they are the magic bullets to fix a system that does not have enough memory. Wrong! This misconception is thought to be true by many people, at various levels, in a lot of organizations. The fact is that if your system does not have enough real memory to run your applications, adding more paging and swap space is not going to help. Depending on the application(s) running on your system, swap space should start at least 1.5 times physical memory. Many high-performance applications require 4 to 6 times real memory so the actual amount of paging and swap space is variable, but 1.5 times is a good place to start. Use the application’s recommended requirement, if one is suggested, as a starting point. Some of you may be asking “What is the difference between paging space and swap space?” It depends on the Unix flavor whether your system does swapping or paging, but both swap space and paging space are disk storage that makes up virtual memory along with real, or physical, memory. A page fault happens when a memory segment, or page, is needed in memory but is not currently resident in memory. When a page fault occurs, the system attempts to load the needed data into memory; this is called paging or swapping, depending on the Unix system you are running. When the system is doing a lot of paging in and out of memory we need to be able to monitor this activity. If your system runs out of paging space or is in a state of continuous swapping, such that as soon as a segment is paged out of memory it is immediately needed again, the

145

Free & Share & Open

146

Chapter 6 system is thrashing. If this thrashing condition continues for very long, you have a risk of the system crashing. In this chapter we are going to use the terms “paging” and “swapping” interchangeably. Each of our four Unix flavors, AIX, HP-UX, Linux, and Solaris, use different commands to list the swap space usage; the output for each command and OS varies also. The goal of this chapter is to create five shell scripts: one script of each of the four operating systems and an all-in-one shell script that will run on any of our four Unix flavors. Each of the shell scripts must produce the exact same output, which is shown in Listing 6.1.
Paging Space Report for yogi Wed Jun 5 21:48:16 EDT 2002 336MB 33MB 303MB 10% 90%

Total MB of Paging Space: Total MB of Paging Space Used: Total MB of Paging Space Free: Percent of Paging Space Used: Percent of Paging Space Free:

Listing 6.1 Required paging and swap space report.

Before we get started creating the shell scripts, we need the command syntax for each operating system. Each of the commands produces a different result, so this should be an interesting chapter in which we can try some varied techniques.

Syntax
As usual, we need the correct command syntax before we can write a shell script. As we go through each of the operating systems, the first thing I want you to notice is the command syntax used and the output received back. Because we want each Unix flavor to produce the same output, as shown in Listing 6.1, we are going to have to do some math. This is not going to be hard math, but each of the paging and swap space command outputs is lacking some of the desired information so we must calculate the missing pieces. Now we are going to see the syntax for each operating system.

AIX lsps Command
AIX does paging instead of swapping. This technique uses 4096-byte blocks pages. When a page fault occurs, AIX has a complex algorithm that frees memory of the least used noncritical memory page to disk paging space. When the memory has space

Monitoring Paging and Swap Space available, the page of data is paged in to memory. To monitor paging space usage in AIX, you use the lsps command, which stands for list paging space. The lsps command has two command options, -a, to list each paging space separately, and -s, to show a summary of all paging spaces. Both lsps options are shown here:
# lsps -a Page Space Physical Volume paging00 hdisk2 hd6 hdisk0 Volume Group Size rootvg 1024MB rootvg 1024MB %Used Active Auto 11 yes yes 9 yes yes Type lv lv

147

# lsps -s Total Paging Space 2048MB Percent Used 10%

From the first command output, lsps -a, on this system notice that there are two paging spaces defined, paging00 and hd6, both are the same size at 1GB each, and each paging space is on a separate disk. This is an important point. In AIX, paging space is used in a round-robin fashion, starting with the paging space that has the largest area of free space. If one paging space is significantly larger, the round-robin technique is defeated, and the system will almost always use the larger paging space. This has a negative effect on performance because one disk will take all of the paging activity. In the second output, lsps -s, we get a summary of all of the paging space usage. Notice that the only data that we get is the total size of the paging space and the percentage used. From these two pieces of data we must calculate the remaining parts of our required output, which is total paging space in MB, free space in MB, used space in MB, percent used, and percent free. We will cover these points in the scripting section for AIX later in this chapter.

HP-UX swapinfo Command
The HP-UX operating system uses swapping, which is evident by the command swapinfo. HP-UX does the best job of giving us the best detailed command output so we need to calculate only one piece of data for our required output, percent of total swap space free. Everything else is provided with the swapinfo -tm command. The -m switch specifies to produce output in MB, and the -t switch specifies to produce a total line for a summary of all virtual memory. This command output is shown here.
[root@dino]/> swapinfo -tm Mb Mb Mb TYPE AVAIL USED FREE dev 96 21 73 reserve 46 -46 memory 15 5 10 total 111 72 37

PCT USED 22% 33% 65%

START/ Mb LIMIT RESERVE 928768 -

PRI 1

NAME /dev/dsk/c0t6d0

-

0

-

Free & Share & Open

148

Chapter 6
Notice in this output that HP-UX splits up virtual memory into three categories: dev, reserve, and memory. For our needs we could use the summary information that is shown in the total line at the bottom. As you can see on the total line, the total virtual memory is 111MB, the system is consuming 72MB of this total, which leaves 37MB of free virtual memory. The fifth column shows that the system is consuming 65 percent of the available virtual memory. This total row is misleading, though, when we are interested only in the swap space usage. The actual swap space usage is located on the dev row of data at the top of the command output. As you can see, we need to calculate only the percent free, which is a simple calculation.

Linux free Command
Linux uses swapping and uses the free command to view memory and swap space usage. The free command has several command switches, but the only one we are concerned with is the -m command switch to list output in MB. The swap information given by the free -m command is listed only in MB, and there are no percentages presented in the output. Therefore, from the total MB, used MB, and free MB, we must calculate the percentages for percentage used and percentage free. The following shows the free -m command output:
# free -m total Mem: 52 -/+ buffers/cache: Swap: 211 used 51 30 9 free 1 22 202 shared 0 buffers 1 cached 20

The last line in this output has the swap information listed in MB, specified by the -m switch. This command output shows that the system has 211MB of total swap space, of which 9MB has been used and 202MB of swap space is free.

Solaris swap Command
The Solaris operating system does swapping, as indicated by the command swap. Of the swap command switches we are concerned with only the -s switch, which produces a summary of swap space usage. All output from this command is produced in KB so we have to do a little division by 1,000 to get our standard MB output. Like Linux, the Solaris swap output does not show the swap status using percentages, so we must calculate these values. The swap -s output is shown here.
# swap -s total: 26788k bytes allocated + 7256k reserved = 34044k used, 557044k available

This is an unusual output to decipher because the data is all on the same line, but because Solaris attempts to create a mathematical statement we will have to use our

Monitoring Paging and Swap Space own mathematical statements to fill in the blanks to get our required script output. The swap -s command output shows that the system has used a total of 34MB and it has 557MB of free swap space. We must calculate the total MB, the percentage used, and the percentage of free swap space. These calculations are not too hard to handle as we will see in the shell scripting section for Solaris later in this chapter.

149

Creating the Shell Scripts
Now that we have the basic syntax of the commands to get paging and swap space statistics, we can start our scripting of the solutions. In each case you should notice which pieces of data are missing from our required output, as shown in Listing 6.1. All of these shell scripts are different. Some pipe command outputs to a while loop to assign the values to variables, and some use other techniques to extract the desired data from the output. Please study each shell script in detail, and you will learn how to handle the different situations you are challenged with when working in a heterogeneous environment.

AIX Paging Monitor
As we previously discussed, the AIX lsps -s command output shows only the total amount of paging space measured in MB and the percentage of paging space that is currently in use. To get our standard set of data to display we need to do a little math. This is not too difficult when you take one step at a time. In this shell script let’s use a file to store the command output data. To refresh your memory the lsps -s command output is shown again here (this output is using a different AIX system):
# lsps -s Total Paging Space 336MB

Percent Used 2%

The first thing we need to do is to remove the columns heading. I like to use the tail command in a pipe for this purpose. The command syntax is shown in the next statement:
# lsps -s | tail +2 336MB

2%

This resulting output contains only the data, without the columns heading. The next step is to store these values in variables so that we can work with them for some calculations. We are going to use a file for initial storage and then use a while read loop, which we feed from the bottom using input redirection with the filename. Of course, we could have piped the command output to the while read loop, but I want to vary the techniques in each shell script in this chapter. Let’s look at the first part of the data gathering and the use of the while read loop, as shown in Listing 6.2.

Free & Share & Open

150

Chapter 6

PAGING_STAT=/tmp/paging_stat.out # Paging Stat hold file # Load the $PAGING_STAT file with data lsps -s | tail +2 > $PAGING_STAT # Use a while loop to assign the values to variables while read TOTAL PERCENT do DO CALCULATIONS HERE done < $PAGING_STAT

Listing 6.2 Logical view of AIX lsps -s data gathering.

Notice in Listing 6.2 that we first define a file to hold the data, which is pointed to by the $PAGING_STAT variable. In the next step we redirect output of our paging space status command to the defined file. Next comes a while loop where we read the file data and assign the first data field to the variable TOTAL and the second data field to the variable PERCENT. Notice how the $PAGING_STAT file is used to feed the while loop from the bottom. As you saw in Chapter 2, “Twelve Ways to Process a File Line by Line,” this technique is one of the two fastest methods of reading data from a file. The middle of the while loop is where we do our calculations to fill in the blanks of our required output. Speaking of calculations, we need to do three calculations for this script, but before we can perform the calculations on the data we currently have, we need to get rid of the suffixes attached to the variable data. The first step is to extract the MB from the $TOTAL variable and then extract the percent sign, %, from the $PERCENT variable. We do both of these operations using a cut command in a pipe, as shown here:
PAGING_MB=$(echo $TOTAL | cut -d ‘MB’ -f1) PAGING_PC=$(echo $PERCENT | cut -d% -f1)

In both of these statements we use command substitution, specified by the $(command_statement) notation, to execute a command statement and assign the result to the variable specified. In the first statement we echo the $TOTAL variable and pipe the output to the cut command. For the cut command we specify the delimiter to be MB, and we enclose it with single tic marks, ‘MB’. Then we specify that we want the first field, specified by -f1. In the second statement we do the exact same thing except that this time we specify that the percent sign, %, is the delimiter. The result of these two statements is that we have the PAGING_MB and PAGING_PC variables pointing to integer values without any other characters. Now we can do our calculations! Let’s do the most intuitive calculation first. We have the value of the percent of paging space used stored in the $PAGING_PC variable as an integer value. To get the

Monitoring Paging and Swap Space percent of free paging space, we need to subtract the percent used value from 100, as shown in the next command statement.
(( PAGING_PC_FREE = 100 - PAGING_PC ))

151

Notice that we used the double parentheses mathematical method, specified by the (( Math Statement )). I like this method because it is so intuitive to use. Also notice that you do NOT use the dollar sign, $, with variables when using this method. Because the double parentheses method expects a mathematical statement, any character string that is not numeric is assumed to be a variable, so the dollar sign should be omitted. If you add a dollar sign to the variable name, then the statement may fail depending on the OS you are running! I always remove the dollar sign, just in case. This is a common cause of frustration when using math in shell scripts, and it is extremely hard to troubleshoot. The next calculation is not so intuitive to some. We want to calculate the MB of paging space that is currently in use. Now let’s think about this. We have the percentage of paging space used, the percentage of paging space free, and the total amount of paging space measured in MB. To calculate the MB of used paging space, we can use the value of the total MB of paging space and the percentage of paging space used divided by 100, which converts the value of paging space used into a decimal value internally. See how this is done in the next statement.
(( MB_USED = PAGING_MB * PAGING_PC / 100 ))

One thing to note in the last math statement: This will produce only an integer output. If you want to see the output in floating-point notation, then you need to use the bc utility, which you will see in some of the following sections. The last calculation is another intuitive calculation, to find the MB of free paging space. Because we already have the values for the total paging space in MB, and the MB of paging space in use, then we need only to subtract the used value from the total. This is shown in the next statement.
(( MB_FREE = PAGING_MB - MB_USED ))

We have completed all of the calculations so now we are ready to produce the required output for the AIX shell script. Take a look at the entire shell script shown in Listing 6.3, and pay particular attention to the boldface type.
#!/usr/bin/ksh # # SCRIPT: AIX_paging_mon.ksh # # AUTHOR: Randy Michael # DATE: 5/31/2002 # REV: 1.1.P #

Listing 6.3 AIX_paging_mon.ksh shell script listing. (continues)

Free & Share & Open

152

Chapter 6

# PLATFORM: AIX Only # # PURPOSE: This shell script is used to produce a report of # the system’s paging space statistics including: # # Total paging space in MB, MB of free paging space, # MB of used paging space, % of paging space used, and # % of paging space free # # REV LIST: # # # set -x # Uncomment to debug this shell script # set -n # Uncomment to check command syntax without any execution # ########################################################### ################ DEFINE VARIABLES HERE #################### PC_LIMIT=65 # Percentage Upper limit of paging space # before notification

THISHOST=$(hostname) # Host name of this machine PAGING_STAT=/tmp/paging_stat.out # Paging Stat hold file ########################################################### ################ INITIALIZE THE REPORT #################### echo “\nPaging Space Report for $THISHOST\n” date ########################################################### ############# CAPTURE AND PROCESS THE DATA ################ # Load the data in a file without the column headings lsps -s | tail +2 > $PAGING_STAT # Start a while loop and feed the loop from the bottom using # the $PAGING_STAT file as redirected input, after “done” while read TOTAL PERCENT do # Clean up the data by removing the suffixes PAGING_MB=$(echo $TOTAL | cut -d ‘MB’ -f1) PAGING_PC=$(echo $PERCENT | cut -d% -f1) # Calculate the missing data: %Free, MB used and MB free (( PAGING_PC_FREE = 100 - PAGING_PC ))

Listing 6.3 AIX_paging_mon.ksh shell script listing. (continued)

Monitoring Paging and Swap Space

153

(( MB_USED = PAGING_MB * PAGING_PC / 100 )) (( MB_FREE = PAGING_MB - MB_USED )) # Produce the rest of the paging space report: echo “\nTotal MB of Paging Space:\t$TOTAL” echo “Total MB of Paging Space Used:\t${MB_USED}MB” echo “Total MB of Paging Space Free:\t${MB_FREE}MB” echo “\nPercent of Paging Space Used:\t${PERCENT}” echo “\nPercent of Paging Space Free:\t${PAGING_PC_FREE}%” # Check for paging space exceeded the predefined limit if ((PC_LIMIT $LOGFILE &

Listing 8.5 proc_watch.ksh shell script listing. (continues)

Free & Share & Open

226

Chapter 8

echo “START PROCESS: $PROCESS began ==> $TIMESTAMP” > $TTY } #################################################### ############## START OF MAIN ####################### #################################################### ### SET A TRAP #### trap ‘trap_exit; exit 0’ 1 2 3 15 # Check for the Correct Command Line Argument - Only 1 if (( $# != 1 )) then usage exit 1 fi # Get an Initial Process State and Set the RUN Flag ps -ef | grep -v “grep $PROCESS” | grep -v $SCRIPT_NAME \ | grep $PROCESS >/dev/null PROC_RC=$? # Check the Return Code!!

# Give some initial feedback before starting the loop if (( PROC_RC == 0 )) then echo “The $PROCESS process is currently running...Monitoring...” RUN=”Y” # Set the RUN Flag to YES else echo “The $PROCESS process is not currently running...Monitoring...” RUN=”N” # Set the RUN Flag to NO fi TIMESTAMP=$(date +%D@%T) # Grab a timestamp for the log # Use a “tee -a $#LOGFILE” to send output to both standard output # and to the file referenced by $LOGFILE echo “MON_START: Monitoring for $PROCESS began ==> $TIMESTAMP” \ | tee -a $LOGFILE # Loop Forever!! while :

Listing 8.5 proc_watch.ksh shell script listing. (continued)

Process Monitoring

227

do case $RUN in ‘Y’) # Loop Until the Process Ends RUN=$(mon_proc_end) ;; ‘N’) # Loop Until the Process Starts RUN=$(mon_proc_start) ;; esac done # End of Script

Listing 8.5 proc_watch.ksh shell script listing. (continued)

The shell script in Listing 8.5 is a nice, modular shell script. The actual monitoring loop is the final while loop. The loop is short and tight, with all of the work being done within the two functions, proc_mon_start and proc_mon_end. Notice that in both functions we remain in the loop until there is a transition from run to stop or not running to process startup. On each transition we return updated run status information back to the calling shell script with a print command, as opposed to a return code. For the concurrent display to the screen and logging to the file we use tee -a $LOGFILE within the shell script body, and in the functions we redirect output to the tty device that we assigned to the $TTY variable. We use the tty device to ensure that the screen output will go to the terminal, or pseudo-terminal, that we are currently looking at. Otherwise we cannot be assured where standard output is pointing within the function. We again did all numeric tests with the double parentheses method. Notice that we do not use a $ with a user-defined variable! For the while loop we are looping forever. The No-Op character (while :) allows this to work (true would also work). The proc_watch.ksh shell script will continue to run until it is interrupted—for example, CTRL-C is pressed. We have improved our script, but it does not let us know how many processes are active. There is no timing mechanism for the shell script; it just runs until interrupted. We are next going to expand on our script to do a few things differently. First, we want to be able to time the monitoring to execute for a specific period of time. We also want to let the user know how many processes are currently active and the PID of each process. In addition, we want to time stamp each process startup and end time. To time stamp each process we can count the number of processes that are running during each loop iteration, and if the count changes we will grab a new time stamp and update the PID list for the currently running processes. We also will give the option to run some pre, startup, and/or post event before the process starts, as the process starts, or after the process has ended.

Free & Share & Open

228

Chapter 8

[root:yogi]@/scripts/WILEY/PROC_MON# ./proc_watch.ksh xcalc The xcalc process is currently running...Monitoring... MON_START: Monitoring for xcalc began ==> 09/27/01@21:09:41 END PROCESS: xcalc ended ==> 09/27/01@21:09:56 START PROCESS: xcalc began ==> 09/27/01@21:10:06 END PROCESS: xcalc ended ==> 09/27/01@21:10:25 ^C MON_STOP: Monitoring for xcalc ended ==> 09/27/01@21:10:31

Listing 8.6 proc_watch.ksh shell script in action.

Timed Execution for Process Monitoring, Showing each PID, and Time Stamp with Event and Timing Capability
Sound like a lot? After we get through this section, each step will be intuitively obvious. In all of the previous three scripts, we had no ability to monitor each process that matched the grepped pattern or to execute the monitoring for a specific amount of time. Because we are using the grep command we may get multiple matches to a pattern. In case of multiple matches we need to know (1) how many matches we have and (2) each process that was matched. This information can be very beneficial if you are monitoring a specific user’s activities or anything where we are interested in the exact process IDs that are running. We also want a good timing mechanism that will allow for easy, flexible timing of the duration of the monitoring activity. Because we have no way of knowing what user requirements may be, we want to allow for as much flexibility as possible. Let’s go to the far side and allow timing from seconds to days, and anything in between. The easiest way to handle timing, but not the most accurate, is to add up all of the seconds and count down from the total seconds to zero while sleeping for one second between counts. We could continuously check the date/time using the date command for a very accurate time, or—even better—we can kick off an at job to kill the script at some specific time in the future. The Korn shell variable SECONDS is also useful. For this script we are going to use getopts to parse the command line for seconds, minutes, hours, days, and the process to monitor. Then we add up the seconds and count down to zero and quit. Alternatively, if the total seconds and a process are the only arguments, the user will be able to enter these directly—for only a process and total seconds getopts will not be used. The usage function will list two ways to use our new script. Another nice option is the capability to run pre, startup, and/or post events. By pre, startup, and post events we are talking about running some command, script, or function before the process starts, as the process starts, or after the process stops, or in any

Process Monitoring combination. As an example, we may want to reboot the machine after a backup program ends, or we may want to set up environment variables before some process starts up. For the event options we also need to be as flexible as possible. For flexibility we will just add a function for each event that contains only the no-op character, : (colon), as a place holder. A colon does not execute anything; it does nothing and has a return code of 0, zero. Anything that a user may want to run before startup, at startup, or after the process has ended can be added into the appropriate function. We will use flags, or variables, to enable and disable the pre, startup, and post events individually. In this section we are going to do two things that may be new, using getopts to process the command-line arguments and executing a function in the background as a co-process. The getopts functionality is an easy and efficient way to parse through mixed command-line arguments, and the command switches can be with or without switch arguments. A co-process is an easy way to set up a communication link with a background script or function and the foreground. Let’s first look at how to use getopts to parse the command line. The getopts command is built in to Korn shell. The command parses the command line for valid options specified by a single character, following a - (minus sign) or a + (plus sign). To specify that a command switch requires an argument, the switch character must be followed by a : (colon). If the switch does not require any argument, then the : should be omitted. All of the switch options put together are called the OptionString, and this is followed by some variable name that we define. The argument for each switch is stored in a variable called $OPTARG as the arguments are parsed in a loop one at a time. If the entire OptionString is preceded by a : (colon), then any unmatched switch option causes a ? to be loaded into the variable that we defined in the getopts command. The form of the command follows: getopts OptionString Name [ Argument ... ]

229

The easiest way to explain the getopts command is with an example. For our script we need seconds, minutes, hours, days, and a process to monitor. For each one we want to supply an argument—for example, -s 5 -m10 -p my_backup. In this example we are specifying 5 seconds, 10 minutes, and the process is my_backup. Notice that there does not have to be a space between the switch and the argument. This is what makes getopts so great! The code to set up our example looks like the following example in Listing 8.7.

SECS=0 MINUTES=0 HOURS=0 DAYS=0 PROCESS=

# Initialize all to zero

# Initialize to null

while getopts “:s:m:h:d:p:” TIMED 2>/dev/null do case $TIMED in s) SECS=$OPTARG

Listing 8.7 Example getopts command usage. (continues)

Free & Share & Open

230

Chapter 8

;; m) (( MINUTES = $OPTARG * 60 )) ;; h) (( HOURS = $OPTARG * 3600 )) ;; d) (( DAYS = $OPTARG * 86400 )) ;; p) PROCESS=$OPTARG ;; \?) usage exit 1 ;; esac done (( TOTAL_SECONDS = SECS + MINUTES + HOURS + DAYS ))

Listing 8.7 Example getopts command usage. (continued)

There are a few things to note in Listing 8.7. The getopts command needs to be part of a while loop with a case statement within the loop. On each option we specified, -s,-m,-h,-d, and -p, and we added a : (colon) after each switch character. This tells getopts that an argument is required for that particular switch character. The : (colon) before the OptionString list tells getopts that if an unspecified option is given on the command line, to set the $TIMED variable to the ? character. The ? allows us to call the usage function and exit with a return code of 1 for an incorrect command-line option. The only thing to be careful of is that getopts does not care what arguments it receives so it is our responsibility to check each argument to ensure that it meets our expectations; then we have to take action if we want to exit. The last thing to note in Listing 8.7 is that the first line of the while loop has redirection of the standard error (file descriptor 2) to the bit bucket. Any time an unexpected argument is encountered, getopts sends a message to standard error, but it is not considered an error, just informational. Because we expect that incorrect command-line arguments may be entered, we can just ignore the messages and discard them with redirection to /dev/null, a.k.a. the bit bucket. We also need to cover setting up a co-process. A co-process is a communications link between a foreground and a background process. The most common question is, “Why is this needed?” In our next script we are going to call a function that will handle all of the monitoring for us while we do the timing control in the main script. The problem arises because we need to run this function in the background. Within the background process monitoring function there are two loops in which one loop is always executing. Without the ability to tell the loop to break out of the internal loop, it will continue to execute on its own after the main script, and function, have exited due to an interrupt. We know what this causes—one or more defunct processes! From the main script we need

Process Monitoring a way to communicate with the loop in the background function to tell it to break out of the loop or exit the function cleanly when the countdown is complete and if the script is interrupted—for example, with CTRL-C. To solve this little problem we kick off our background proc_watch function as a co-process. “How do we do this?” you ask. “Pipe it to the background” is the simplest way to put it, and that is also what it looks like. Look at the next example in Listing 8.8. function proc_watch { # This function is started as a co-process!!! while : # Loop forever do Some Code Here read BREAK_OUT # Do NOT need a “-p” to read! if [[ $BREAK_OUT = ‘Y’ ]] then return 0 fi done } ############################ ##### Start of Main ######## ############################ ### Set a Trap ### trap ‘BREAK=’Y’; print -p $BREAK; exit 2’ 1 2 3 15 TOTAL_SECONDS=300 BREAK_OUT=’N’ proc_watch |& # Start proc_watch as a co-process!!!!

231

until (( TOTAL_SECONDS == 0 )) do (( TOTAL_SECONDs = TOTAL_SECONDS - 1 )) sleep 1 done BREAK_OUT=’Y’ # Use “print -p” to communicate with the co-process variable print -p $BREAK_OUT exit 0

Listing 8.8 Example using a co-process.

Free & Share & Open

232

Chapter 8
In the code block in Listing 8.8 we defined the proc_watch function, which is the function that we want to start as a background process. As you can see, the proc_watch function has an infinite loop. If the main script, is interrupted, then without a means to exit the loop within the proc_watch background function, the loop alone will continue to execute! To solve this we start the proc_watch as a co-process by “piping it to the background” using pipe ampersand, |&, as a suffix. Now when we want to communicate with the function from the main script, we use print -p $BREAK_OUT. Inside the function we just use the standard read command, read BREAK_OUT. The co-process is the mechanism that we are going to use to break out of the loop if the main script is interrupted on a trapped signal, and for normal countdown termination at the end of the script. Of course, we can never catch kill -9 with a trap. Try setting up the scenario just described, without a co-process, with a background function that has an infinite loop. Then press the CTRL-C key sequence to kill the main script and do a ps -ef | more. You will see that the background loop is still executing! Get the PID, and do a kill -9 to kill it. Of course, if the loop’s exit criteria is ever met, the loop will exit on its own. Now take a look at the entire script, and see how we handled all of these extra requirements. Pay close attention to the highlighted code in Listing 8.9.

#!/bin/ksh # # SCRIPT: proc_watch_timed.ksh # AUTHOR: Randy Michael # DATE: 09-14-2001 # REV: 1.0.P # PLATFORM: Not Platform Dependent # # PURPOSE: This script is used to monitor and log # the status of a process as it starts and stops. # Command line options are used to identify the target # process to monitor and the length of time to monitor. # Each event is logged to the file defined by the # $LOGFILE variable. This script also has the ability # to execute pre, startup, and post events. These are # controlled by the $RUN_PRE_EVENT, $RUN_STARTUP_EVENT, and # $RUN_POST_EVENT variables. These variables control execution # individually. Whatever is to be executed is to be placed in # either the “pre_event_script”, startup_event_script, or the # “post_event_script” functions, or in any combination. Timing # is controlled on the command line. # # USAGE: $SCRIPT_NAME total_seconds target_process # # Will monitor the specified process for the

Listing 8.9 proc_watch_timed.ksh shell script listing.

Process Monitoring

233

# specified number of seconds. # # USAGE: $SCRIPT_NAME [-s|-S seconds] [-m|-M minutes] # [-h|-H hours] [-d|-D days] # [-p|-P process] # # Will monitor the specified process for number of # seconds specified within -s seconds, -m minutes, # -h hours, and -d days. Any combination of command # switches can be used. # # REV LIST: # # set -x # Uncomment to debug this script # set -n # Uncomment to check syntax without ANY execution # #################################################### ########## DEFINE FILES AND VARIABLES HERE ######### #################################################### typeset -u RUN_PRE_EVENT # Force to UPPERCASE typeset -u RUN_STARTUP_EVENT # Force to UPPERCASE typeset -u RUN_POST_EVENT # force to UPPERCASE RUN_PRE_EVENT=’N’ # A ‘Y’ will execute, anything else will not RUN_STARTUP_EVENT=’Y’ # A ‘Y’ will execute, anything else will not RUN_POST_EVENT=’Y’ # A ‘Y’ will execute, anything else will not LOGFILE=”/tmp/proc_status.log” [[ ! -s $LOGFILE ]] && touch $LOGFILE SCRIPT_NAME=$(basename $0) TTY=$(tty) INTERVAL=”1” # Seconds between sampling JOBS= #################################################### ############# DEFINE FUNCTIONS HERE ################ #################################################### usage () { echo “\n\n\t*****USAGE ERROR*****” echo “\n\nUSAGE: $SCRIPT_NAME seconds process” echo “\nWill monitor the specified process for the” echo “specified number of seconds.” echo “\nUSAGE: $SCRIPT_NAME [-s|-S seconds] [-m|-M minutes]” echo “ [-h|-H hours] [-d|-D days] [-p|-P process]\n”

Listing 8.9 proc_watch_timed.ksh shell script listing. (continues)

Free & Share & Open

234

Chapter 8

echo “\nWill monitor the specified process for number of” echo “seconds specified within -s seconds, -m minutes,” echo “-h hours and -d days. Any combination of command” echo “switches can be used.\n” echo “\nEXAMPLE: $SCRIPT_NAME 300 dtcalc” echo “\n\nEXAMPLE: $SCRIPT_NAME -m 5 -p dtcalc” echo “\nBoth examples will monitor the dtcalc process” echo “for 5 minutes. Can specify days, hours, minutes” echo “and seconds, using -d, -h, -m and -s\n\n” } #################################################### trap_exit () { # set -x # Uncommant to debug this function # Log an ending time for process monitoring echo “INTERRUPT: Program Received an Interrupt...EXITING...” > $TTY echo “INTERRUPT: Program Received an Interrupt...EXITING...” >> $LOGFILE TIMESTAMP=$(date +%D@%T) # Get a new time stamp... echo “MON_STOPPED: Monitoring for $PROCESS ended ==> $TIMESTAMP\n” \ >> $TTY echo “MON_STOPPED: Monitoring for $PROCESS ended ==> $TIMESTAMP\n” \ >> $LOGFILE echo “LOGFILE: All Events are Logged ==> $LOGFILE \n” > $TTY # Kill all functions JOBS=$(jobs -p) if [[ ! -z $JOBS && $JOBS != ‘’ && $JOBS != ‘0’ ]] then kill $(jobs -p) 2>/dev/null 1>&2 fi return 2 } #################################################### pre_event_script () { # Put anything that you want to execute BEFORE the # monitored process STARTS in this function : # No-OP - Needed as a place holder for an empty function # Comment Out the Above colon, ‘:’ PRE_RC=$? return $PRE_RC } #################################################### startup_event_script () {

Listing 8.9 proc_watch_timed.ksh shell script listing. (continued)

Process Monitoring

235

# Put anything that you want to execute WHEN, or AS, the # monitored process STARTS in this function : # No-OP - Needed as a place holder for an empty function # Comment Out the Above colon, ‘:’ STARTUP_RC=$? return $STARTUP_RC } #################################################### post_event_script () { # Put anything that you want to execute AFTER the # monitored process ENDS in this function : # No-OP - Need as a place holder for an empty function # Comment Out the Above colon, ‘:’ POST_RC=$? return $POST_RC } #################################################### # This function is used to test character strings test_string () { if (( $# != 1 )) then echo ‘ERROR’ return fi C_STRING=$1 # Test the character string for its composition case $C_STRING in +([0-9])) echo ;; +([-0-9])) echo ;; +([a-z])) echo ;; +([A-Z])) echo ;; +([a-z]|[A-Z])) ;; ‘POS_INT’ # Integer >= 0 ‘NEG_INT’ # Integer < 0 ‘LOW_CASE’ ‘UP_CASE’ # lower case text # UPPER case text

echo ‘MIX_CASE’ # MIxed CAse text

Listing 8.9 proc_watch_timed.ksh shell script listing. (continues)

Free & Share & Open

236

Chapter 8

*) echo ‘UNKNOWN’ # Anything else esac } #################################################### proc_watch () { # set -x # Uncomment to debug this function # This function does all of the process monitoring! while : # Loop Forever!! do case $RUN in ‘Y’) # This will run the startup_event_script, which is a function if [[ $RUN_STARTUP_EVENT = ‘Y’ ]] then echo “STARTUP EVENT: Executing Startup Event Script...”\ > $TTY echo “STARTUP EVENT: Executing Startup Event Script...”\ >> $LOGFILE startup_event_script # USER DEFINED FUNCTION!!! RC=$? # Check the Return Code!! if (( “RC” == 0 )) then echo “SUCCESS: Startup Event Script Completed RC ${RC}” > $TTY echo “SUCCESS: Startup Event Script Completed RC ${RC}” >> $LOGFILE else echo “FAILURE: Startup Event Script FAILED RC ${RC}” > $TTY echo “FAILURE: Startup Event Script FAILED RC ${RC}” >> $LOGFILE fi fi integer PROC_COUNT=’-1’ # Reset the Counters integer LAST_COUNT=’-1’ # Loop until the process(es) end(s) until (( “PROC_COUNT” == 0 )) do # This function is a Co-Process. $BREAK checks to see if # “Program Interrupt” has taken place. If so BREAK will # be ‘Y’ and we exit both the loop and function. read BREAK

Listing 8.9 proc_watch_timed.ksh shell script listing. (continued)

Process Monitoring

237

if [[ $BREAK = ‘Y’ ]] then return 3 fi PROC_COUNT=$(ps -ef | grep -v “grep $PROCESS” \ | grep -v $SCRIPT_NAME \ | grep $PROCESS | wc -l) >/dev/null 2>&1 if (( “LAST_COUNT” > 0 && “LAST_COUNT” != “PROC_COUNT” )) then # The Process Count has Changed... TIMESTAMP=$(date +%D@%T) # Get a list of the PID of all of the processes PID_LIST=$(ps -ef | grep -v “grep $PROCESS” \ | grep -v $SCRIPT_NAME \ | grep $PROCESS | awk ‘{print $2}’) echo “PROCESS COUNT: $PROC_COUNT $PROCESS\ Processes Running ==> $TIMESTAMP” >> $LOGFILE & echo “PROCESS COUNT: $PROC_COUNT $PROCESS\ Processes Running ==> $TIMESTAMP” > $TTY echo ACTIVE PIDS: $PID_LIST >> $LOGFILE & echo ACTIVE PIDS: $PID_LIST > $TTY fi LAST_COUNT=$PROC_COUNT sleep $INTERVAL # Needed to reduce CPU load! done RUN=’N’ # Turn the RUN Flag Off TIMESTAMP=$(date +%D@%T) echo “ENDING PROCESS: $PROCESS END time $TIMESTAMP” >> $LOGFILE & echo “ENDING PROCESS: $PROCESS END time $TIMESTAMP” > $TTY

==>\ ==>\

# This will run the post_event_script, which is a function if [[ $RUN_POST_EVENT then echo “POST EVENT: > $TTY echo “POST EVENT: >> $LOGFILE = ‘Y’ ]] Executing Post Event Script...”\ Executing Post Event Script...”\ &

post_event_script # USER DEFINED FUNCTION!!! integer RC=$? if (( “RC” == 0 ))

Listing 8.9 proc_watch_timed.ksh shell script listing. (continues)

Free & Share & Open

238

Chapter 8

then echo “SUCCESS: Post Event Script Completed RC ${RC}” > $TTY echo “SUCCESS: Post Event Script Completed RC ${RC}” >> $LOGFILE else echo “FAILURE: Post Event Script FAILED RC - ${RC}”\ > $TTY echo “FAILURE: Post Event Script FAILED RC - ${RC}”\ >> $LOGFILE fi fi ;; ‘N’) # This will run the pre_event_script, which is a function if [[ $RUN_PRE_EVENT = ‘Y’ ]] then echo “PRE EVENT: Executing Pre Event Script...” > $TTY echo “PRE EVENT: Executing Pre Event Script...” >> $LOGFILE pre_event_script # USER DEFINED FUNCTION!!! RC=$? # Check the Return Code!!! if (( “RC” == 0 )) then echo “SUCCESS: Pre Event Script Completed > $TTY echo “SUCCESS: Pre Event Script Completed >> $LOGFILE else echo “FAILURE: Pre Event Script FAILED RC > $TTY echo “FAILURE: Pre Event Script FAILED RC >> $LOGFILE fi fi echo “WAITING: Waiting for $PROCESS to startup...Monitoring...” integer PROC_COUNT=’-1’ # Initialize to a fake value # Loop until at least one process starts until (( “PROC_COUNT” > 0 ))

RC - ${RC}”\ RC - ${RC}”\

- ${RC}”\ - ${RC}”\

Listing 8.9 proc_watch_timed.ksh shell script listing. (continued)

Process Monitoring

239

do # This is a Co-Process. This checks to see if a “Program # Interrupt” has taken place. If so BREAK will be ‘Y’ and # we exit both the loop and function read BREAK if [[ $BREAK = ‘Y’ ]] then return 3 fi PROC_COUNT=$(ps -ef | grep -v “grep $PROCESS” \ | grep -v $SCRIPT_NAME | grep $PROCESS | wc -l) \ >/dev/null 2>&1 sleep $INTERVAL # Needed to reduce CPU load! done RUN=’Y’ # Turn the RUN Flag On TIMESTAMP=$(date +%D@%T) PID_LIST=$(ps -ef | grep -v “grep $PROCESS” \ | grep -v $SCRIPT_NAME \ | grep $PROCESS | awk ‘{print $2}’) if (( “PROC_COUNT” == 1 )) then echo “START PROCESS: $PROCESS START time ==> $TIMESTAMP” >> $LOGFILE & echo ACTIVE PIDS: $PID_LIST >> $LOGFILE & echo “START PROCESS: $PROCESS START time ==> $TIMESTAMP” > $TTY echo ACTIVE PIDS: $PID_LIST > $TTY elif (( “PROC_COUNT” > 1 )) then echo “START PROCESS: $PROC_COUNT $PROCESS Processes Started: START time ==> $TIMESTAMP” >> $LOGFILE & echo ACTIVE PIDS: $PID_LIST >> $LOGFILE & echo “START PROCESS: $PROC_COUNT $PROCESS Processes Started: START time ==> $TIMESTAMP” > $TTY echo ACTIVE PIDS: $PID_LIST > $TTY fi ;; esac done

Listing 8.9 proc_watch_timed.ksh shell script listing. (continues)

Free & Share & Open

240

Chapter 8

} #################################################### ############## START OF MAIN ####################### #################################################### ### SET A TRAP #### trap ‘BREAK=’Y’;print -p $BREAK 2>/dev/null;trap_exit\ 2>/dev/null;exit 0’ 1 2 3 15 BREAK=’N’ # The BREAK variable is used in the co-process proc_watch PROCESS= # Initialize to null integer TOTAL_SECONDS=0 # Check commnand line arguments if (( $# > 10 || $# < 2 )) then usage exit 1 fi # Check to see if only the seconds and a process are # the only arguments if [[ ($# -eq 2) && ($1 != -*) && ($2 != -*) ]] then NUM_TEST=$(test_string $1) # Is this an Integer? if [[ “$NUM_TEST” = ‘POS_INT’ ]] then TOTAL_SECONDS=$1 # Yep - It’s an Integer PROCESS=$2 # Can be anything else usage exit 1 fi else # Since getopts does not care what arguments it gets lets # do a quick sanity check to make sure that we only have # between 2 and 10 arguments and the first one must start # with a -* (hyphen and anything), else usage error case “$#” in [2-10]) if [[ $1 != -* ]]; then usage; exit 1 fi ;;

Listing 8.9 proc_watch_timed.ksh shell script listing. (continued)

Process Monitoring

241

esac HOURS=0 # Initialize all to zero MINUTES=0 SECS=0 DAYS=0 # Use getopts to parse the command line arguments # For each $OPTARG for DAYS, HOURS, MINUTES and DAYS check to see # that each one is an integer by using the check_string function while getopts “:h:H:m:M:s:S:d:D:P:p:” OPT_LIST 2>/dev/null do case $OPT_LIST in h|H) [[ $(test_string $OPTARG) != ‘POS_INT’ ]] && usage && exit (( HOURS = $OPTARG * 3600 )) # 3600 seconds per hour ;; m|H) [[ $(test_string $OPTARG) != ‘POS_INT’ ]] && usage && exit (( MINUTES = $OPTARG * 60 )) # 60 seconds per minute ;; s|S) [[ $(test_string $OPTARG) != ‘POS_INT’ ]] && usage && exit SECS=”$OPTARG” # seconds are seconds ;; d|D) [[ $(test_string $OPTARG) != ‘POS_INT’ ]] && usage && exit (( DAYS = $OPTARG * 86400 )) # 86400 seconds per day ;; p|P) PROCESS=$OPTARG # process can be anything ;; \?) usage # USAGE ERROR exit 1 ;; :) usage exit 1 ;; *) usage exit 1 ;; esac done fi # We need to make sure that we have a process that # is NOT null or empty! - sanity check - The double quotes are required! if [[ -z “$PROCESS” || “$PROCESS” = ‘’ ]] then usage

1

1

1

1

Listing 8.9 proc_watch_timed.ksh shell script listing. (continues)

Free & Share & Open

242

Chapter 8

exit 1 fi # Check to see that TOTAL_SECONDS was not previously set if (( TOTAL_SECONDS == 0 )) then # Add everything together if anything is > 0 if [[ $SECS -gt 0 || $MINUTES -gt 0 || $HOURS -gt 0 \ || $DAYS -gt 0 ]] then (( TOTAL_SECONDS = SECS + MINUTES + HOURS + DAYS )) fi fi # Last Sanity Check! if (( TOTAL_SECONDS $TTY ps -ef | grep -v “grep $PROCESS” | grep -v $SCRIPT_NAME \ | grep $PROCESS > $TTY PROC_RC=$? # Get the initial state of the monitored function echo >$TTY # Send a blank line to the screen (( PROC_RC != 0 )) && echo “\nThere are no $PROCESS processes running\n” if (( PROC_RC == 0 )) # The Target Process(es) is/are running... then RUN=’Y’ # Set the RUN flag to true, or yes. integer PROC_COUNT # Strips out the “padding” for display PROC_COUNT=$(ps -ef | grep -v “grep $PROCESS” | grep -v \ $SCRIPT_NAME | grep $PROCESS | wc -l) >/dev/null 2>&1 if (( PROC_COUNT == 1 ))

Listing 8.9 proc_watch_timed.ksh shell script listing. (continued)

Process Monitoring

243

then echo “The $PROCESS process is currently running...Monitoring...\n” elif (( PROC_COUNT > 1 )) then print “There are $PROC_COUNT $PROCESS processes currently running...Monitoring...\n” fi else echo “The $PROCESS process is not currently running...monitoring...” RUN=’N’ # Set the RUN flag to false, or no. fi TIMESTAMP=$(date +%D@%T) # Time that this script started monitoring # Get a list of the currently active process IDs PID_LIST=$(ps -ef | grep -v “grep $PROCESS” \ | grep -v $SCRIPT_NAME \ | grep $PROCESS | awk ‘{print $2}’) echo “MON_STARTED: Monitoring for $PROCESS began ==> $TIMESTAMP” \ | tee -a $LOGFILE echo ACTIVE PIDS: $PID_LIST | tee -a $LOGFILE

##### NOTICE #### # We kick off the “proc_watch” function below as a “Co-Process” # This sets up a two way communications link between the # “proc_watch” background function and this “MAIN BODY” of # the script. This is needed because the function has two # “infinite loops”, with one always executing at any given time. # Therefore we need a way to break out of the loop in case of # an interrupt, i.e. CTRL-C, and when the countdown is complete. # The “pipe appersand”, |&, creates the background Co-Process # and we use “print -p $VARIABLE” to transfer the variable’s # value back to the background co-process. ################################### proc_watch |& WATCH_PID=$! # Create a Background Co-Process!! # Get the process ID of the last background job!

# Start the Count Down! integer SECONDS_LEFT=$TOTAL_SECONDS while (( SECONDS_LEFT > 0 )) do

Listing 8.9 proc_watch_timed.ksh shell script listing. (continues)

Free & Share & Open

244

Chapter 8

# Next send the current value of $BREAK to the Co-Process # proc_watch, which was piped to the background... print -p $BREAK 2>/dev/null (( SECONDS_LEFT = SECONDS_LEFT - 1 )) sleep 1 # 1 Second Between Counts done # Finished - Normal Timeout Exit... TIMESTAMP=$(date +%D@%T) # Get a new time stamp... echo “MON_STOPPED: Monitoring for $PROCESS ended ==> $TIMESTAMP\n” \ | tee -a $LOGFILE echo “LOGFILE: All Events are Logged ==> $LOGFILE \n” # Tell the proc_watch function to break out of the loop and die BREAK=’Y’ print -p $BREAK 2>/dev/null kill $WATCH_PID 2>/dev/null exit 0 # End of Script

Listing 8.9 proc_watch_timed.ksh shell script listing. (continued)

The most important things to note in Listing 8.9 are the communication link used between the foreground main script and the background co-process function, proc_watch, and the use of getopts to parse the command-line arguments. Some other things to look at are the integer tests using the string_test function and the way that the user is notified of a new process either starting or stopping by time stamp. The updated process count and the listing of all of the PIDs and how text is sent to the tty display within the function. As usual, we use the double parentheses numerical test method in the control structures. (Notice again that the $ is not used to reference the user defined variables!) This shell script is also full of good practices for using different control structures and the use of the logical AND and logical OR (&& and ||), which reduces the need for if..then..else.. and case structures. One very important test needs to be pointed out—the “null/empty” test for the PROCESS variable just after getopts parses the command line. This test is so important because the getopts command does not care what arguments it is parsing;, nothing will “error out.” For this reason, we need to verify all of the variables ourselves. The only thing getopts is doing is

Process Monitoring matching the command switches to the appropriate arguments, not the validity of the command-line argument! If this test is left out and invalid command line arguments are present, then grep command errors will cover the screen during the script’s execution— bad, very bad! A good review of Listing 8.9 is needed to point out some other interesting aspects. Let’s start at the top: In the definitions of the files and variables there are three variables that control the execution of the pre, startup, and post events. The variables are RUN_PRE_EVENT, RUN_STARTUP_EVENT, and RUN_POST_EVENT, and for ease of testing, the variables are typeset to UPPERCASE. A 'Y' will enable the execution of the function, in which a user can put anything that he or she wants to run. The functions are called pre_event_script, startup_event_script, and post_event_script, but don’t let the names fool you. We also identify the LOGFILE variable and test to see if a log file exists. If the file does not exist, we touch the $LOGFILE variable, which creates an empty file with the filename that the $LOGFILE variable points to. This script section also grabs the SCRIPT_NAME using the basename $0 command, and we define the current tty device for display purposes. An important variable is INTERVAL. This variable defines the number of seconds between sampling the process list. It is very important that this value is greater than 0, zero! If the INTERVAL value is set to 0, zero, then the CPU load will be extreme and will produce a noticeable load, to say the least. The next section in Listing 8.9 defines all of the functions used in this script. We have a usage function that is displayed for usage errors. Then there is the trap_exit function. The trap_exit function will execute on exit codes 1, 2, 3, and 15, which we will see in the trap statement later at “Start of Main” in the script. Next are the pre_event_script, startup_event_script, and post_event_script functions. You may ask why a function would have a name indicating it is a script. It is done this way to encourage the use of an external script, or program, for any pre, startup, or post event activity, rather than editing this script and debugging an internal function. The next function is used to test character strings, thus the name test_string. If you have ever wondered how to test a string (the entire string!) for its composition, test_string will do the trick. We just use a regular expression test for a range of characters. The preceding + (plus sign) is required in this regular expression to specify that all characters are of the specified type. Then comes the main function in the script that does all of the work, proc_watch. This function is also the one that is executed as the co-process that we have been talking so much about. The proc_watch function is an infinite loop that contains two internal loops, where one internal loop is always executing at any given time. During both of these internal loops we check the variable BREAK to see if the value is 'Y'. The 'Y' value indicates that the function should exit immediately. The BREAK variable is updated, or changed, from the main script and is “transferred” to this co-process background function using the print -p $BREAK command within the main script. This variable is reread, in the function, on each loop iteration using the standard read BREAK command. This is what enables the clean exit from the background function’s loop. The word background is key to understanding the need for the co-process. If the main script is interrupted, then the innermost loop will continue to execute even after both the function and script end execution. It will exit on its own when the loop’s exit

245

Free & Share & Open

246

Chapter 8 parameters are met, but if they are never met we end up with a defunct process. To get around this problem we start the proc_watch function as a background co-process using |& as a suffix to the function—for example, proc_watch |&. An easy way to think of a co-process is a pipe to the background, and through this pipe we have a communications link. For the main part of the shell script, at the START OF MAIN, we first set a trap. In the trap command we set the BREAK variable to 'Y', to indicate that the proc_watch co-process should exit, and we make the new BREAK value known to the co-process with the print -p $BREAK 2>/dev/null command. This command sometimes sends error notification to the standard error, file descriptor 2, but we want all error notification suppressed. Otherwise, the error messages would go to the screen during the script’s execution, which is highly undesirable. Next are the standard things of initializing a few variables and checking for the correct number of arguments. There are two ways to run this script: (1) only specifying the total seconds and the process to monitor or (2) using the command-line switches to specify the seconds, minutes, hours, days, and process to monitor. The latter method will use the getopts command to parse the arguments, but we do not need getopts for the first method. We first check to see if we are given only seconds and a process. We use the test_string function to ensure that the $1 argument is a positive integer. The second argument could be anything except a string that begins with a - (hyphen) or a null string. Otherwise, we will use the getopts command to parse the command line. Using the getopts command makes life much easier when we need to process command-line arguments; however, getopts does have its limitations. The command is parsing the command-line arguments, but it really does not care what the arguments are. Therefore, we need to do a sanity check on each and every argument to ensure that it meets the criteria that is expected. If the argument fails, then we just run the usage function and exit with a return code of 1, one. Two tests are conducted on each argument. We test the PROCESS variable to make sure that it is not null, or empty, and we check all of the numeric variables used for timing to make sure they are positive integers, or 0, zero. The positive integer test is to ensure that at least one of the numeric variables, SECS, MINUTES, HOURS, and DAYS, has an integer value greater than 0, zero. If we get past this stage we assume we have creditable data to start monitoring. The monitoring starts by getting an initial state of the process, either currently running or not running. With this information we initialize the RUN variable, which is used as a control mechanism for the rest of the script. Once the initialization text is both logged and sent to the screen, the proc_watch function is started as a background coprocess, again using proc_watch |&. The main script just does a countdown to 0, zero, and exits. To make the proc_watch function exit cleanly we assign 'Y' to the BREAK variable and make this new value known to the co-process with the print -p $BREAK command. Then we kill the background PID that we saved in the WATCH_PID variable and then exit the script with a return code of 0, zero. If the script is interrupted, then the trap will handle stopping the co-process and exiting. See Listing 8.10.

Process Monitoring

247

[root:yogi]@/scripts/WILEY/PROC_MON# ./proc_watch_timed.ksh -m 5 -pxcalc Currently running xcalc processes: There are no xcalc processes running The xcalc process is not currently running...monitoring... MON_STARTED: Monitoring for xcalc began ==> 09/27/01@21:15:02 ACTIVE PIDS: START PROCESS: xcalc START time ==> 09/27/01@21:15:19 ACTIVE PIDS: 26190 STARTUP EVENT: Executing Startup Event Script... SUCCESS: Startup Event Script Completed RC - 0 PROCESS COUNT: 2 xcalc Processes Running ==> 09/27/01@21:15:46 ACTIVE PIDS: 13060 26190 PROCESS COUNT: 3 xcalc Processes Running ==> 09/27/01@21:16:04 ACTIVE PIDS: 13060 18462 26190 PROCESS COUNT: 4 xcalc Processes Running ==> 09/27/01@21:16:27 ACTIVE PIDS: 13060 18462 22996 26190 PROCESS COUNT: 3 xcalc Processes Running ==> 09/27/01@21:16:39 ACTIVE PIDS: 18462 22996 26190 PROCESS COUNT: 4 xcalc Processes Running ==> 09/27/01@21:16:56 ACTIVE PIDS: 18462 22996 24134 26190 PROCESS COUNT: 3 xcalc Processes Running ==> 09/27/01@21:17:31 ACTIVE PIDS: 22996 24134 26190 PROCESS COUNT: 2 xcalc Processes Running ==> 09/27/01@21:17:41 ACTIVE PIDS: 22996 24134 PROCESS COUNT: 3 xcalc Processes Running ==> 09/27/01@21:18:39 ACTIVE PIDS: 21622 22996 24134 PROCESS COUNT: 2 xcalc Processes Running ==> 09/27/01@21:18:58 ACTIVE PIDS: 21622 22996 PROCESS COUNT: 3 xcalc Processes Running ==> 09/27/01@21:19:04 ACTIVE PIDS: 18180 21622 22996 PROCESS COUNT: 4 xcalc Processes Running ==> 09/27/01@21:19:10 ACTIVE PIDS: 18180 21622 22758 22996 PROCESS COUNT: 6 xcalc Processes Running ==> 09/27/01@21:19:17 ACTIVE PIDS: 18180 21622 22758 22996 23164 26244 PROCESS COUNT: 5 xcalc Processes Running ==> 09/27/01@21:19:37 ACTIVE PIDS: 18180 22758 22996 23164 26244 PROCESS COUNT: 4 xcalc Processes Running ==> 09/27/01@21:19:47 ACTIVE PIDS: 18180 22996 23164 26244 PROCESS COUNT: 3 xcalc Processes Running ==> 09/27/01@21:19:53 ACTIVE PIDS: 18180 22996 26244

Listing 8.10 proc_watch_times.ksh shell script in action. (continues)

Free & Share & Open

248

Chapter 8

PROCESS COUNT: 2 xcalc Processes Running ==> 09/27/01@21:19:55 ACTIVE PIDS: 18180 26244 PROCESS COUNT: 1 xcalc Processes Running ==> 09/27/01@21:20:05 ACTIVE PIDS: 18180 PROCESS COUNT: 0 xcalc Processes Running ==> 09/27/01@21:20:09 ACTIVE PIDS: ENDING PROCESS: xcalc END time ==> 09/27/01@21:20:11 POST EVENT: Executing Post Event Script... SUCCESS: Post Event Script Completed RC - 0 MON_STOPPED: Monitoring for xcalc ended ==> 09/27/01@21:20:23 LOGFILE: All Events are Logged ==> /tmp/proc_status.log

Listing 8.10 proc_watch_times.ksh shell script in action. (continued)

Other Options to Consider
The proc_watch_timed.ksh shell script is thorough, but it may need to be tailored to a more specific need. Some additional considerations are listed next.

Common Uses
These scripts are suited for things like monitoring how long a process runs, logging a process as it starts and stops, restarting a process that has terminated prematurely, and monitoring a problem user or contractor. We can also monitor activity on a particular tty port and send an email as a process starts execution. Use your imagination. We can start the monitoring script on the command line, or as a cron or at job, and run it during the work day. A cron table entry might look like the following:
0 7 * * 1-5 /usr/local/bin/proc_watch_timed.ksh -h9 -p fred >/dev/null

This cron table entry would monitor any process in the process table that contained “fred” from 7:00 A.M. Monday through Friday for nine hours. Note: The nine hours may be much longer due to the system’s load during the day as the script counts down to zero. Anything in the system’s process list can be monitored from seconds to days.

Modifications to Consider
These scripts are generic, and you may want to make modifications. One option to consider is to list the actual lines in the process list instead of only the PID and a process

Process Monitoring count with a time stamp. You may have a different ps command that is preferred—for example, ps aux. For a more accurate timing you may want to check the date/time at longer intervals (as opposed to counting down); checking the time would also reduce the CPU load. Another good idea is to get the timing data and run an at command to kill the script at the specified time. Also, consider using the Korn shell built-in variable SECONDS. First initialize the SECONDS variable to 0, zero and it will automatically increment each second as long as the parent process is executing. The pre, startup, and post events are something else to look at, the startup in particular. The startup_ event_script currently executes only when (1) the monitoring starts and the target process is running and (2) when the very first, if more than one, process starts, not as each process starts. You may want to modify this function’s execution to run only as each individual process starts and not to execute when monitoring starts and the target process is already running. Additionally, depending on what is to be executed for any of these events, some sleep time might be needed to allow for things to settle down. As we can see, there are many ways to do all of this, and everyone has different expectations and requirements. Just remember that we never have a final script; we just try to be flexible!

249

Summary
In this chapter we started with a very basic idea of monitoring for a process to start or stop. We quickly built on user options to monitor the process state for a specified period of time and added time stamps. We also allowed the user to specify pre, startup, and post events to execute as an option. Never try to do everything at once. Build a short shell script that does the basic steps of your target goal and expand on the base shell script to build in the nice-to-have things. I use the proc_mon.ksh and proc_wait.ksh shell scripts almost daily for monitoring system events and they sure do save a lot of time reentering the same command over-and-over again. In the next chapter we are going to expand on our monitoring to include applications. We always want to know if an application or database has gone down during the day. As you watch the heads popping up above all of the cubicles it is always nice to be proactively informed and not be surprised by the application group.

Free & Share & Open

CHAPTER

9
Monitoring Processes and Applications

The most critical part of any business is ensuring that applications continue to run without error. In this chapter, we are going to look at several techniques for monitoring applications and critical processes that the applications rely on. The problem with trying to write this chapter is that there are so many applications in the corporate world that the techniques to monitor them vary widely. From the lowest level we can ping the machine to see if it is up. A ping, though, is not an operating system response, but rather a machine response to confirm that the network adapter is configured. At a higher level, we can look at the processes that are required for the application to run properly, but this too does not completely confirm, 100 percent, that the application is working properly. The only way to ensure the application is working properly is to interact with the application. As an example, if we have a database that the application requires we can do a simple SQL query to ensure that the database is working properly. For interactive applications we can try to use a here document to log in to the application and maybe even perform a small task. Applications work differently, so the solution to ensure that the application is up and running properly will vary widely. We are going to look at monitoring local processes, remote monitoring using Secure Shell (SSH), checking for active Oracle databases, and checking an application URL and HTTP server status in this chapter.

251

Free & Share & Open

252

Chapter 9

Monitoring Local Processes
Above pinging a host machine, the most common application monitoring technique is to look for the critical processes that are required for the application to work properly. This is also a good practice when we have a flaky application that has a process that dies intermittently. The basic technique is to use the ps -ef | grep target_process | grep -v grep command syntax. If you have more than one required process, then this command statement needs to be executed for each of the processes individually. We do not want to use egrep in place of grep in this case. If egrep is used, then we get a positive result if any of the processes are currently running. The key to making this technique work is to find a unique string pattern that represents the target process. The PID is no good because the process may have a child or parent process that has the same PID somewhere in the ps -ef output. Finding a unique string pattern that works with the grep command is key. This is easily tested by using the following command syntax on the command line. ps -ef | grep Appserver | grep -v grep

This command statement assumes that we are looking for a process called Appserver. Notice that we always pipe (|) the last pipe’s output to | grep -v grep. This last grep on grep is needed so that the system will not report on the grep Appserver process. In the process table each part of the command statement that has a pipe will have a separate PID. Then there is another thing to consider if this command is executed in a shell script. The shell script name may show up in the grep output, depending on how the shell script is written. To get around this little problem we need to query the system to capture the shell script’s filename and add a third grep to the ps -ef command statement using the following syntax:
SCRIPT_NAME=$(basename $0) ps -ef | grep target_process | grep -v grep | grep -v $SCRIPT_NAME

Now we have a command that will work if, and only if, a unique character string can be found that separates the target process from all other processes. This usually takes a few tries for each application that we want to monitor. In Listing 9.1 we have a code segment from a shell script that monitors an application service, using a unique character string. This particular application service is defined by the APPSVC variable. If this service is not currently running, there is an attempt to restart the application service and an email is sent to my text pager and my regular email account. Follow the code segment in Listing 9.1.

Monitoring Processes and Applications

253

###################################################### ############# DEFINE VARIABLES HERE ################## ###################################################### APPSVC=”/usr/local/sbin/appstrt_u1” MAILLIST=”1234567890@mypage.provider.abc randy@my.domain.com” MAILFILE=”/tmp/mailfile.out” TIMESTAMP=$(date +%m%d%y%H%M%S) APPS_LOG=”/usr/local/log/appsvc.log [ -s $APP_LOG ] || touch $APP_LOG ###################################################### ################# START OF MAIN ###################### ###################################################### # Check to see if the APPSVC process(es) is/are running APPSVC_COUNT=$(ps -ef | grep $APPSVC | grep -v grep \ | grep -v $SCRIPT_NAME | wc -l) # If the count is zero then we need to attempt to restart the service if (( $APPSVC_COUNT == 0 )) then # Need to attempt an Application server restart. echo “SVC-A1 - APPSVC: Attempting Restart” > $MAILFILE # Send email notification sendmail -f rmichael@my.domain.com $MAILLIST < $MAILFILE # Make a log entry echo “ERROR: $TIMESTAMP - Appsvc DOWN - Attempting Restart”>> $APP_LOG # Make another log entry echo “STARTING APPLICATION SERVER - $TIMESTAMP” >>$APPS_LOG # Attempt the restart!!! su - appsvc -c ‘/usr/local/sbin/appsvc start 2>&1’ >> $APP_LOG fi

Listing 9.1 Code segment to monitor an application process.

In the code segment in Listing 9.1 notice that we defined a unique string for the process, which in this case is the fully qualified pathname, to the APPSVC variable. Because this application server can have multiple instances running at the same time,

Free & Share & Open

254

Chapter 9 we need to get a count of how many of these processes are running. If the process count is 0, zero, a restart of the application server is attempted. During the restart effort an email notification is sent to reflect that the application service is down and the script is attempting a restart. This information is also logged in the $APP_LOG file before the restart command. Notice the restart command at the end of the script segment. This monitoring script is executed from the root crontab every 10 minutes. Because the script is running as root it is easy to use the su (switch user) command to execute a single command as the appsvc user for the restart. If you are not familiar with this technique, then study the syntax in Listing 9.1 and study the man page for the su command.

Remote Monitoring with Secure Shell
In the previous section we studied a “local” shell script. No one said, though, that you could not run this same script from a remote machine. This is where Open Secure Shell (OpenSSH) comes into play. Open Secure Shell is a freeware encryption replacement for telnet, ftp, and rsh, for the most part. When we use the ssh command we establish a connection between two machines, and a secure tunnel allows encrypted communication between two trusted machines. Using ssh we can log in to another trusted machine in the network, we can copy files between the machines in an encrypted state, and we can run commands on a remote trusted machine. OpenSSH can be downloaded at the following URL: http://www.openssh.com To establish password-free encrypted connections, an encryption key pair must be created on both machines. This encryption key is located on both machines in the user’s $HOME/.ssh directory. All of the details to set up the password-free encrypted connections are shown in great detail in the ssh man page (man ssh). Let’s look at a couple of examples of using ssh. The first example shown in Listing 9.2 shows a simple login without the key pair created.
# ssh randy@dino The authenticity of host ‘dino (10.10.10.6)’ can’t be established. RSA key fingerprint is c5:19:37:b9:59:ad:3a:18:6b:45:57:2d:ab:b8:df:bb. Are you sure you want to continue connecting (yes/no)? yes Warning: Permanently added ‘dino (10.10.10.6)’ (RSA) to the list of known hosts. randy@dino’s password: Last unsuccessful login: Tue Jul 2 13:58:18 EDT 2002 on /dev/pts/24 from bambam

Listing 9.2 Sample secure shell login.

Monitoring Processes and Applications

255

Last login: Wed Aug

7 10:28:00 EDT 2002 on /dev/pts/18 from bambam

************************************************************************ * * * Welcome to dino! * * * Please see the README file in /usr/lpp/bos for information pertinent * to this release of the AIX Operating System. * * * ************************************************************************ [YOU HAVE NEW MAIL] [randy@dino] $

Listing 9.2 Sample secure shell login. (continued)

Notice in Listing 9.2 that the login to dino required a password, which indicates that the systems do not have the encryption key pairs set up. This does get a bit annoying when you are trying to run a command on a remote machine using an ssh tunnel. With the key pairs created on both machines we can monitor remote machines using encryption, and no password is required. As an example, suppose I need to check the filesystem usage on dino and I am logged into yogi. By adding the command that we want to execute on dino to the end of the ssh login statement, we establish a trusted connection between the two machines, and the command executes on the remote machine with the output going to the local machine. Of course, this is equivalent to a remote shell, rsh, except that the information is encrypted using ssh in place of rsh. A simple example of this technique is shown in Listing 9.3.

[randy@yogi] ssh randy@dino df -k Filesystem /dev/hd4 /dev/hd2 /dev/hd9var /dev/hd3 1024-blocks 196608 1441792 2162688 131072 Free %Used 66180 67% 488152 67% 1508508 31% 106508 19% Iused %Iused Mounted on 2330 3% / 29024 9% /usr 868 1% /var 361 2% /tmp

Listing 9.3 Example of running a remote command. (continues)

Free & Share & Open

256

Chapter 9

/dev/hd1 /dev/local_lv /dev/oracle_lvx /dev/arch_lvx

589824 235556 61% 393216 81384 80% 1507328 307388 80% 13631488 8983464 35%

15123 2971 5008 44

11% /home 4% /usr/local 2% /oracle 1% /oradata

Listing 9.3 Example of running a remote command. (continued)

Notice in the output in Listing 9.3 that there was no prompt for a password and that the result was presented back to the local terminal. Once the key pairs are set up you can do remote monitoring with ease, as long as your security staff does not find any bugs in the ssh code. Let’s move on to Oracle now.

Checking for Active Oracle Databases
I wanted to have at least one example of interacting with an application in this chapter, and I picked an Oracle database as the example using a SQL+ database query. We will look at three steps to check the Oracle database status. The first step is to list all of the Oracle instances defined in the /etc/oratab file. This file is colon-separated (:) with the Oracle instance name(s) in the first field, $1. The function shown in Listing 9.4 first checks to see if a /etc/oratab file exists. If the file is not found, then a notification message is displayed on the user’s terminal and the function returns a 3 for a return code. Otherwise, the /etc/oratab file is parsed to find the Oracle instance name(s). Removing all of the lines that begin with comments, specified by beginning with a hash mark (#), in the file is done using a sed statement in combination with the ^# notation. Removing the comment lines is easy using the sed statement, as shown here with a /etc/hosts file as an example. cat /etc/hosts | sed /^#/b > /etc/hosts.without_beginning_comments

The output of the previous command shows all of the IP address and hostname entries, except that the commented-out lines have been removed. The ^# is the key to finding the commented lines, which translates to begins with a #. Check out the function in Listing 9.4 to see how we use this technique to parse the Oracle instances from the /etc/oratab file. function show_oratab_instances { if [ ! -f “$ORATAB” ] then echo “\nOracle instance file $ORATAB does not exist\n” return 3 else cat $ORATAB | sed /^#/b | awk -F: ‘{print $1}’ fi }

Listing 9.4 show_oratab_instances function listing.

Monitoring Processes and Applications
The output of the show_oratab_instances function in Listing 9.4 is a list of all of the Oracle instances defined on the system. We have already removed the lines that are comments; next comes the awk statement that extracts the first field, specified by awk -F: ‘{print $1}’. In this awk statement the -F: specifies that the line is field separated by colons (:). Once we know the field separator we just extract the first field ($1), which is the Oracle instance name. Now we are going to use the same function shown in Listing 9.4 to get the status of all of the defined Oracle instances by checking for the process for each instance. This technique is shown in Listing 9.5. function show_all_instances_status { for INSTANCE in $(show_oratab_instances) do ps -ef | grep ora | grep $INSTANCE | grep -v grep >/dev/null 2>&1 if (($? != 0)) then echo “\n$INSTANCE is NOT currently running $(date)\n” else echo “\n$INSTANCE is currently running OK $(date)\n” fi done }

257

Listing 9.5 show_all_instances_status function listing.

Notice in Listing 9.5 that we use the function from Listing 9.4 to get the list of Oracle instances to query the system for. In this case, all we are doing is using the ps -ef command again. This time we narrow the list down with a grep on the string ora. This output is piped (|) to another grep statement, where we are looking for the instance name for the current loop iteration, specified by $INSTANCE. Of course, we need to strip out any grep processes from the output so we add one more grep -v grep. If the return code of the entire ps -ef statement is 0, zero, then the instance is running; if the return is anything other than 0, zero, then the instance is not running. We are still looking at the process level. I have seen cases when the instance processes are running, but I still could not log in to the database. For a final test we need to do an actual SQL query of the database to interact with Oracle. This just needs to be a very simple query to prove that we can interact with the database and get data back. To actually query the Oracle database we can use a simple SQL+ statement, as shown in Listing 9.6. This two-line SQL script is used in the function simple_ SQL_query, shown in Listing 9.6 using the sqlplus command. select * from user_users; exit

Listing 9.6 my_sql_query.sql SQL script listing.

Free & Share & Open

258

Chapter 9
As you can see in Listing 9.6, this is not much of a query, but it is all that we need. This SQL script, my_sql_query.sql, is used in the sqlplus function in Listing 9.7. Notice in this function, simple_SQL_query, that the sqlplus command statement requires a username, password, and an Oracle SID name to work. See the function code in Listing 9.7. function simple_SQL_query { USER=oracle PASSWD=oracle SID=yogidb sqlplus ${USER}/${PASSWD}@$SID @my_sql_query.sql }

Listing 9.7 simple_SQL_query function listing.

The function shown in Listing 9.7 can be shortened further, if you are logged in to the system as the oracle user or executing a script as the oracle user. If these conditions are met then you can run a simpler version of the previous sqlplus, as shown in Listing 9.8, with the output of the query; however, the Oracle Listener is not tested as in the previous sqlplus statement in Listing 9.7. The sqlplus command in Listing 9.8 should be run on the local machine.
[oracle@yogi] sqlplus / @/usr/local/bin/mysql_query.sql SQL*Plus: Release 8.1.7.0.0 - Production on Wed Aug 7 16:07:30 2002 (c) Copyright 2000 Oracle Corporation. All rights reserved.

Connected to: Oracle8i Enterprise Edition Release 8.1.7.4.0 - Production With the Partitioning option JServer Release 8.1.7.4.0 - Production

USERNAME USER_ID ACCOUNT_STATUS ----------------------------- ---------- -------------------------------LOCK_DATE EXPIRY_DATE DEFAULT_TABLESPACE ----------- ----------- -----------------------------TEMPORARY_TABLESPACE CREATED INITIAL_RSRC_CONSUMER_GROUP

Listing 9.8 Example of an SQL+ Oracle query.

Monitoring Processes and Applications

259

------------------------------ ----------- -----------------------------EXTERNAL_NAME --------------------------------------------------------------------------OPS$ORACLE 940 OPEN USERS TEMP 18-APR-2002

Disconnected from Oracle8i Enterprise Edition Release 8.1.7.4.0 Production With the Partitioning option JServer Release 8.1.7.4.0 - Production

Listing 9.8 Example of an SQL+ Oracle query. (continued)

This is about as simple as it gets! You can check the return code from the sqlplus command shown in Listing 9.8. If it is zero, then the query worked. If the return code is nonzero, then the query failed and the database should be considered down. In any case, the Database Administrator needs to be notified of this condition.

Checking If the HTTP Server/Application Is Working
Some applications use a Web browser interface. For this type of application we can use a command-line browser, such as linx, to attempt to reach a specific URL, which in turn should bring up the specified application Web page. The function shown in Listing 9.9 utilizes the linx command-line browser to check both the HTTP server and the Web page presented by the specified URL, which is passed to the function in the $1 argument. check_HTTP_server () { LINX=”/usr/local/bin/lynx” URL=$1 URLFILE=/tmp/HTTP.$$

# Define the location of the linx program # Capture the target URL in the $1 position # Define a file to hold the URL output

########################################### $LINX “$URL” > $URLFILE if (($? != 0)) # Attempt to reach the target URL # If the URL is unreachable - No Connection

Listing 9.9 check_HTTP_server function listing. (continues)

Free & Share & Open

260

Chapter 9

then echo “\n$URL - Unable to connect\n” cat $URLFILE else # Else the URL was found while read VER RC STATUS redirection do case $RC in 200|401|301|302) # This while loop is fed from the bottom # after the “done” using input

# Check the return code in the $URLFILE # These are valid return codes! echo “\nHTTP Server is OK\n” ;; # Anything else is not a valid return code echo “\nERROR: HTTP Server Error\n” ;;

*)

esac done < $URLFILE fi rm -f $URLFILE }

Listing 9.9 check_HTTP_server function listing. (continued)

This is a nice function in Listing 9.9 for checking the status of a Web server and also to see if an application URL is accessible. You should test this function against doing the same task manually using a graphical browser. This has been tested on an application front-end, and it works as expected; however, a good test is recommended before implementing this, or any other code, in this book. You know all about the disclaimer stuff. (I am really not even here writing this book, or so the disclaimer says.)

Other Things to Consider
As with any code that is written, it can always be improved. Each of the functions and code segments presented in this chapter are just that, code segments. When you are monitoring applications, code like this is only one part of a much bigger shell script, at least it should be. The monitoring should start at the lowest level, which is sending a ping to the application host to ensure that the machine is powered on and booted. Then we apply more layers as we try to build a script that will allow us to debug the problem. I have presented only a few ideas; it is your job to work out the details for your environment.

Monitoring Processes and Applications

261

Application APIs and SNMP Traps
Most enterprise management tools come with application program interfaces (APIs) for the more common commercial applications; however, we sometimes must write shell scripts to fill in the gaps. This is where SNMP traps come in. Because the enterprise management tool should support SNMP traps, the APIs allow the application to be monitored using the SNMP MIB definitions on both the management server and the client system. When an enterprise management tool supports SNMP traps, you can usually write your own shell scripts that can use the tool’s MIB and SNMP definitions to get the message out from your own shell scripts. As an example, the command shown here utilizes a well-known monitoring tool’s SNMP and MIB data to allow a trap to be sent.
/usr/local/bin/trapclient $MON_HOST $MIB_NUM $TRAP_NUM $TRAP_TEXT

In the previous command the MON_HOST variable represents the enterprise management workstation. The MIB_NUM variable represents the specific code for the MIB parameter. The TRAP_NUM variable represents the specific trap code to send, and the TRAP_TEXT is the text that is sent with the trap. This type of usage varies depending on the monitoring tool that you are using. At any rate, there are techniques that allow you to write shell scripts to send traps. The methods vary, but the basic syntax remains the same for SNMP.

Summary
This is one of those chapters where it is useless to write a bunch of shell scripts. I tried to show some of the techniques of monitoring applications and application processes, but the details are too varied to cover in a single chapter. I have laid down a specific process that you can utilize to build a very nice tool to monitor your systems and applications. Always start with a ping! If the box is unpingable, then your first job is to get the machine booted or to call hardware support. In the next steps you have several options, including interacting with the application, as we did with a SQL+ query of an Oracle database. We also covered monitoring specific processes that are a little flaky and die every once in a while. I have two applications that I have to monitor this way, and I have not had even one phone call since I put this tool in place. The key is to keep the business in business, and the best way to do that is to be very proactive. This is where good monitoring and control shell scripts make you look like gold. Remember, no one ever notices an application except when it is down! In the next chapter, we move on to study creating pseudo-random passwords. The scripts include the use of arrays in shell scripts and a practical use for computergenerated pseudo-random numbers in a shell script. See you in the next chapter!

Free & Share & Open

CHAPTER

10
Creating Pseudo-Random Passwords

Got security? Most of the user community does not know how to create secure passwords that are not easy to guess. Users tend to have several passwords that they rotate. The problem with these “rotating” passwords is that they are usually easy to guess. For example, users find that birth dates, social security numbers, addresses, department names/numbers, and so on make good passwords that are easy to remember. Sometimes they even use words found in any dictionary, which is a starting point for any cracker. In this chapter we are going to create a shell script that creates pseudo-random passwords.

Randomness
If you look at Chapter 21, “ Pseudo-Random Number Generator,” you can see the exercise that we used to create pseudo-random numbers. These numbers are not true random numbers because of the cyclical nature of how “random numbers” are created on a computer system. For example, if you always start a random number sequence with the same seed, or first number, you will always have the same sequence of numbers. In Chapter 21 we used the process ID (PID) of the current process, which was the shell script, as the seed for creating pseudo-random numbers. This use of the PID is good because PIDs are created by the system in a somewhat random nature. Now that I have lost you in random numbers you are asking, “What does a random number have to do with a password?” As we proceed, the answer will be intuitively obvious.

263

Free & Share & Open

264

Chapter 10

Creating Pseudo-Random Passwords
We started this chapter with a discussion on randomness because we are going to use computer-generated pseudo-random numbers, then use these generated numbers as pointers to specific array elements of keyboard characters, which are stored in the array KEYS. In this chapter you get a practical use for generating random numbers, and you thought Chapter 21 was a waste of time! The script idea goes like this: We use an external file that contains keyboard characters, one character per line. You can put any keyboard characters in this file that you want. I just went down the rows on the keyboard from left to right, starting on the top row of keys with numbers. As I went through all of the keyboard keys I then added a second set of numbers from the number keypad, as well as all of the uppercase and lowercase characters. The nice thing about this strategy is that you have the ability to specify the exact group of characters that make a valid password in your shop. Country-specific keyboards, which use characters other than those of the U.S. keyboards, also benefit from this strategy. Once we have the keyboard file created, we load the keyboard data into an array. Don’t panic! Korn shell arrays are easy to work with, as you will see in the scripting section as well as in the array introduction section. When we have all of the array elements loaded, then we know how many total elements we have to work with. Using techniques described in Chapter 21, we create pseudo-random numbers between one and the total number of array elements, n. With an array pointer, which is nothing more than a pseudo-random number, pointing to an individual character, we add the specific character to build a text string. The default length of this character string, which is the password we are creating, is eight characters; however, this can be changed on the command line to make the password longer or shorter by adding an integer value specifying the new password length. The final step is to print the password to the screen. We also add two command-line switch options, -n and -m. The -n switch specifies that the user wants to create a new keyboard data file. The -m switch specifies that the user wants to print a password page. In our shop we are required to put some passwords, such as root, in multiple security envelopes to be locked in a safe, just in case. To remove the risk of typos, I print the password page, which has three copies of the password data on the same page, and cut the sheet up into three pieces. I then fold each of the three slips of paper and seal each one in a security envelope and give them to my Admin Manager. As you can see, creating passwords is not something that I take lightly! Weak passwords make for a real security risk, and as a Systems Administrator you need to take a proactive approach to create secure passwords that are as random as you can make them. This chapter is a valuable asset to any security team as well as for the common user.

Syntax
As with any of our scripting sessions we first need the correct syntax for the primary commands that we are going to use in the shell script. In this case we need to introduce

Creating Pseudo-Random Passwords arrays and the commands that are used to work with the array and the array elements. There is a lot more than loading an array to creating this shell script. When we get to the scripting section you will see the other tasks that I have in mind, and you can pick up a pointer or two from the chapter.

265

Arrays
In a Korn shell we can create one-dimensional arrays. A one-dimensional array contains a sequence of array elements, which are like the boxcars connected together on a train track. An array element can be just about anything, except for another array. I know, you’re thinking that you can use an array to access an array to create two- and three-dimensional arrays. If this can be done, it is beyond the scope of this book. For our task we are going to load our array with single-character array elements that are loaded into the array from an external file. An array element can be a text string, number, line of text, print queue name, or just about anything you can list.

Loading an Array
An array can be loaded in two ways. You can define and load the array in one step with the set -A command, or you can load the array one element at a time. Both techniques are shown here. Defining and Loading Array “KEYS” in One Step set -A KEYS q w e r t y u i o p \[ \] a s d f g h j k l \$

Notice in this preceding list that the characters [, ], and $ have been escaped to remove their special function by adding a backslash character. If we had not escaped these characters, then errors, and strange behavior, may occur as you tried to load or display the array elements. You will see this on a larger scale in the shell script. Also remember that if you enclose a list in double quotes or single tic marks it is treated as a single array element, not as individual array elements. Loading Array “KEYS” One Array Element at a Time The second option for loading the array KEYS is to use a while read loop and use a file as input to the while loop. In this example we load the array elements one at a time using a counter to index the KEYS array.
X=0 while read ARRAY_ELEMENT do ((X = X + 1)) KEYS[$X]=$ARRAY_ELEMENT done < $ARRAY_ELEMENT_FILE

Free & Share & Open

266

Chapter 10
The first loading option, which uses the set -A command, requires that you hardcode the keyboard layout into the shell script, which removes much of the flexibility that you want when restricting or expanding password content. Using the while loop method we can use an external file and load this file with any characters that we want, and we can have as many or as few characters defined for passwords as we like. We can also duplicate characters and change the order of the characters any way we wish. As the counter is incremented on each while loop iteration, we load the array elements in sequential order, starting with array elements 1, KEYS[1]. When we get to the end of the file, we know how many elements we have loaded in the array by the value of the array counter, $X. To see the specific value of array element 22, you can use the following syntax:
# echo ${KEYS[22]} ;

As you can see from the response, the 22nd array element that was loaded is a semicolon character (;). We can also display the number of array elements using either of the following two options:
# echo ${#KEYS[*]) # echo ${#KEYS[@])

Notice that we started with array element 1, one. The Korn shell also supports array element 0, zero, but the pseudo-random numbers we create start with one, not zero. We will look at arrays more closely as we write our shell script.

Building the Password Creation Script
I want to explain this shell script one step at a time, and we have a lot to cover, so let’s get started. First, you need to understand the order of execution and each task that is involved in this script.

Order of Appearance
As usual, we start out by defining the variables that are required for this script. The following section shows the variables that are defined for this shell script.

Define Variables
LENGTH=8 # Default password length. NOTIFICATION_LIST= # Persons to notify if the password is revealed or the “glass has been broken.” DEFAULT_PRINTER= the password report. # Default printer to print

Creating Pseudo-Random Passwords
SCRIPT=$(basename $0) path removed. # The name of this shell script with the directory # Temporary hold file for the printer report. # File containing keyboard

267

OUTFILE=/tmp/tmppwd.out

KEYBOARD_FILE=/scripts/keyboard.keys characters.

PRINT_PASSWORD_MANAGER_REPORT= # Print report flag. RANDOM=$$ # Initializes the random seed to the PID of the shell script, which is pretty random. The purpose of each of these variables is shown after the pound sign (#) on each line.

Define Functions
We have six functions to go through in this section. The functions described here are listed in their order of appearance in the shell script, mk_passwd.ksh. In each of the function descriptions there is a function listing for you to follow through. in_range_random_number Function Description The Korn shell provides an environment variable called—you guessed it—RANDOM. This pseudo-random number generator uses a seed as a starting point to create all future numbers in the sequence. The initial seed is used to create a pseudo-random number. This resulting number is used for the next seed to create the next random number, and so on. As you would expect, if you always start generating your numbers with the same seed each time, you will get the exact same number sequence each time. To change the repeatability we need to have a mechanism to vary the initial seed each time we start generating numbers. I like to use the current process ID (PID) of the shell script because this number will vary widely and is an easy way to change the seed value each time we start generating numbers. We often want to limit the range of numbers not to exceed a user-defined maximum. An example is creating lottery numbers between 1 and the maximum number, which might be 36. We are going to use the modulo arithmetic operator to reduce all numbers to a fixed set of numbers between [0..N-1], which is called modulo N arithmetic. We are going to use this pseudo-random number to index array elements in the KEYS array. For our number range we need a script-defined maximum value, which we will assign to a variable called UPPER_LIMIT. This UPPER_LIMIT variable is defined when the KEYS array has been loaded because it represents the total number of elements that are contained in the KEYS array. The modulo operator is the percent sign (%), and we use this operator the same way that you use the forward slash (/) in division. We still use the RANDOM Korn shell variable to get a new pseudo-random number. This time, though, we are going to use the following equation to limit the number to not exceed the script-defined maximum.
RANDOM_NUMBER=$(($RANDOM % $UPPER_LIMIT + 1))

Free & Share & Open

268

Chapter 10
Notice that we added one to the result. Using the preceding equation will produce a pseudo-random number between 1 and the script-defined $UPPER_LIMIT, which is the total number of elements in the KEYS array. The function using this equation is in_range_random_number and is shown in Listing 10.1. function in_range_random_number { # Create a pseudo-random number less than or equal # to the $UPPER_LIMIT value, which is defined in the # main body of the shell script. RANDOM_NUMBER=$(($RANDOM % $UPPER_LIMIT + 1)) echo “$RANDOM_NUMBER” }

Listing 10.1 in_range_random_number function listing.

The function in Listing 10.1 assumes that the RANDOM variable seed has been initialized in the main body of the shell script and that a script-defined UPPER_LIMIT variable has been set. This function will produce numbers between 1 and the script-defined maximum value. load_default_keyboard Function Description As it turns out, you can add as many, or as few, characters to the $KEYBOARD_FILE file. What if the user wants a quick startup and an easy way to create this required file? This is the reason why I added this function to the mk_passwd.ksh shell script. There are two mechanisms for loading a default keyboard layout. The first way is when the shell script is unable to locate the $KEYBOARD_FILE on the system. In this case the user is prompted to load the default keyboard layout. The second option is to add -n as a command-line switch. We will get to parsing command-line switches later in this chapter. In either of the two situations the user is still prompted before the $KEYBOARD_FILE is loaded with default keyboard layout. Other than prompting the user to load the default keyboard layout, we need to supply a list of keyboard characters to load into the file. At this point let’s look at the function code in Listing 10.2 and cover the details at the end. function load_default_keyboard { # If a keyboard data file does not exist then the user # is prompted to load the standard keyboard data into the # $KEYBOARD_FILE, which is defined in the main body of

Listing 10.2 load_default_keyboard function listing.

Creating Pseudo-Random Passwords

269

# the shell script. clear # Clear the screen

echo “\nLoad the default keyboard data file? (Y/N): \c” read REPLY case $REPLY in y|Y) : ;; *) echo “\nSkipping the load of the default keyboard file...\n” return ;; esac cat /dev/null > $KEYBOARD_FILE echo “\nLoading the Standard Keyboard File...\c” # Loop through each character in the following list and # append each character to the $KEYBOARD_FILE file. This # produces a file with one character on each line. for CHAR in \` 1 2 3 4 5 6 7 8 9 0 \- \= \\ q w e r t y u i o \ p \[ \] a s d f g h j k l \; \’ z x c v b n m \, \ \. \/ \\ \~ \! \@ \# \$ \% \^ \& \* \( \) _ \+ \| \ Q W E R T Y U I O P \{ \} A S D F G H J K L \: \” \ Z X C V B N M \< \> \? \| \. 0 1 2 3 4 5 6 7 8 9 \/ \ \* \- \+ do echo “$CHAR” >> $KEYBOARD_FILE done echo “\n\n\t...Done...\n” sleep 1

Listing 10.2 load_default_keyboard function listing. (continued)

Now I want to direct your attention to the for loop in Listing 10.2, which is in boldface text. The idea is to loop through each character one at a time and append the character to the $KEYBOARD_FILE. The result is a file that contains the keyboard layout, listed one character per line. The file shows one character per line to make it easier to load the file and the KEYS array. In the list of characters please notice that most of the nonalphanumeric characters are preceded by a backslash (\), not just the Korn shell special characters. As we discussed previously, this backslash is used to escape the special meaning of these characters.

Free & Share & Open

270

Chapter 10
When you precede a special character with the backslash, you are able to use the character as a literal character, just like the alphanumeric characters, and if a backslash precedes the other non-alphanumeric characters, it is ignored. The list of characters that are escaped is shown here:
` ! @ # $ % ^ & * ( ) _ - = + [ ] { }

On each loop iteration one character is appended to the $KEYBOARD_FILE using the following command: echo “$CHAR” >> $KEYBOARD_FILE

When the file is loaded, which happens extremely fast, we notify the user that the load is complete and then sleep for one second. I added this sleep 1 at the end of this function because the load happened so fast that the user needed a second to see the message. check_for_and_create_keyboard_file Function Description Is this function name descriptive enough? I like to know exactly what a function is used for by reading the name of the function. The purpose of this function is to check for the existence of the $KEYBOARD_FILE and to prompt the user to load the default keyboard layout into the $KEYBOARD_FILE. The user has the option to load the default data or not to load it. If the user declines to load the keyboard data file, then this script will not work. To get around this little problem, we just notify the user of this ERROR and exit the shell script. When the user gets the error message, he or she is also informed of the name of the missing file and a description of what the script expects in the file—specifically, one keyboard character per line. The full function is shown in Listing 10.3. function check_for_and_create_keyboard_file { # If the $KEYBOARD_FILE does not exist then # ask the user to load the “standard” keyboard # layout, which is done with the load_default_keyboard # function. if [ ! -s $KEYBOARD_FILE ] then echo “\n\nERROR: Missing Keyboard File” echo “\n\nWould You Like to Load the” echo “Default Keyboard Layout?” echo “\n\t(Y/N): \c” typeset -u REPLY=FALSE read REPLY if [[ $REPLY != Y ]] then

Listing 10.3 check_for_and_create_keyboard_file function listing.

Creating Pseudo-Random Passwords

271

echo echo echo echo echo exit else

“\n\nERROR: This shell script cannot operate” “without a keyboard data file located in” “\n==> $KEYBOARD_FILE\n” “\nThis file expects one character per line.” “\n\t...EXITING...\n” 3

load_default_keyboard echo “\nPress ENTER when you are you ready to continue: \c” read REPLY clear fi fi }

Listing 10.3 check_for_and_create_keyboard_file function listing. (continued)

To check for the existence of the $KEYBOARD_FILE, we use the -s test in an if statement, an shown here: if [ ! -s $KEYBOARD_FILE ] then ... fi

Notice that we negated the test by adding an exclamation point ( ! -s ). This is actually a test to see if the file is not greater than zero bytes in size or that the $KEYBOARD_FILE does not exist. If either of these conditions is met, then we display some messages to the user and ask the user if the default keyboard layout should be loaded. If the user acknowledges the question with a “Y” or a “y,” then we execute the load_default_keyboard function, which we studied in the last section, “load_default_keyboard Function Description.” After the keyboard data is loaded into the $KEYBOARD_FILE, we stop and ask the user to press ENTER to continue. Once the user presses ENTER, the script creates a pseudo-random password, which we will cover in a later section. build_manager_password_report Function Description You may be asking, “Why do you want to print a password?” There are a lot of reasons to print a password, but only one of the answers is valid! For security reasons. Now, I really lost you! How can a printed password be good for security? It’s simple: The root password needs to be protected at all costs. Our machines do not have direct login access to root, but we use an auditing script that captures every keystroke of the root user. If a machine has failed and you need to log on to the system on the console, you are definitely going to need access to the root password. For this reason we keep three copies of the root password in secure envelopes, and they get locked up for safe keeping.

Free & Share & Open

272

Chapter 10
The build_manager_password_report function creates a file, pointed to by the $OUTFILE variable, that has three copies of the same information on a single page. Look at the function shown in Listing 10.4 to see the message. function build_manager_password_report { # Build a file to print for the secure envelope ( echo “\n RESTRICTED USE!!!” echo “\n\n\tImmediately send an e-mail to:\n” echo “ $NOTIFICATION_LIST”

echo “\n\tif this password is revealed!” echo “\n\tAIX root password: $PW\n” echo “\n\n” echo “\n RESTRICTED USE!!!” echo “\n\n\tImmediately send an e-mail to:\n” echo “ $NOTIFICATION_LIST”

echo “\n\tif this password is revealed!” echo “\n\tAIX root password: $PW\n” echo “\n\n” echo “\n RESTRICTED USE!!!” echo “\n\n\tImmediately send an e-mail to:\n” echo “ $NOTIFICATION_LIST”

echo “\n\tif this password is revealed!” echo “\n\tAIX root password: $PW\n” ) > $OUTFILE }

Listing 10.4 build_manager_password_report function listing.

Notice that the entire message is enclosed in parentheses, with the final output redirected to the $OUTFILE file using the following syntax:
( echo statements.... ) > $OUTFILE

Creating Pseudo-Random Passwords
This method runs all of the echo commands as a separate shell and sends the resulting output to the $OUTFILE using output redirection. Also notice the $NOTIFICATION_LIST variable. This variable is set in the main body of the script. This variable contains the list of people who must be notified if the password is ever released, as stated in the message in the function. When I get one of these printouts, I always run to get it as soon as the page comes out of the printer. This is an extremely important piece of paper! I take it to my desk and cut the page into three pieces and seal each one in a secure envelope and have it locked up for safe keeping. A sample manager’s password report is shown in Listing 10.5.
RESTRICTED USE!!!

273

Immediately send an e-mail to: Donald Duck, Yogi Bear, and Mr. Ranger if this password is revealed! AIX root password: E-,6Kc11

RESTRICTED USE!!!

Immediately send an e-mail to: Donald Duck, Yogi Bear, and Mr. Ranger Immediately send an e-mail to: Donald Duck, Yogi Bear, and Mr. Ranger if this password is revealed! AIX root password: E-,6Kc11

RESTRICTED USE!!!

Listing 10.5 Password report printout. (continues)

Free & Share & Open

274

Chapter 10

Immediately send an e-mail to: Donald Duck, Yogi Bear, and Mr. Ranger if this password is revealed! AIX root password: E-,6Kc11

Listing 10.5 Password report printout. (continued)

You need to edit this function and change the message to suit your environment. If you do not need this functionality, then never use the -m switch, or reply “No” when asked to confirm the printing. usage Function Description It is always a good idea to show the user a USAGE: statement when incorrect or insufficient input is detected (we will get to detecting input errors later in this chapter). For our mk_passwd.ksh shell script we have four options and several combinations. We can execute the mk_passwd.ksh script with no arguments, and you can execute the mk_passwd.ksh shell script with the -n and -m command-line switches. The -n switch loads the default keyboard layout into the $KEYBOARD_FILE file. We can also change the length of the password, which is defined as eight characters by default. Any combination of these command options can be executed. Please look closely at the USAGE: statement shown in Listing 10.6.

function usage { echo “\nUSAGE: $SCRIPT [-m] [-n] echo “ Where: -m -n

[password_length]\n”

Creates a password printout for Security Loads the default keyboard data keys file Integer value that overrides the default 8 character password length.\n”

password_length

}

Listing 10.6 usage function listing.

Creating Pseudo-Random Passwords
When a usage error is detected, the script executes the usage function that displays the following message:
USAGE: $SCRIPT [-m] [-n] Where: -m -n [password_length]

275

Creates a password printout for Security Loads the default keyboard data keys file Integer value that overrides the default 8 character password length.\n”

password_length

trap_exit Function Description This function, trap_exit, is executed only when an exit signal is trapped. You will see how to set a trap a little later. The purpose of this function is to execute any command(s) that are listed in the function. In our case, we want to remove the $OUTFILE before exiting the shell script. Additionally, we do not want to see any messages sent to stderr if the file does not exist. The statement is shown in the following code. function trap_exit { rm -f $OUTFILE >/dev/null 2>&1 }

Notice that we redirect the stderr output to stdout, which is specified by the 2>&1 notation, but not before we send everything to the bit bucket, specified by >/dev/null. That is it for the functions. The next section covers the testing and parsing required for the command arguments.

Testing and Parsing Command-Line Arguments
Because this shell script has command-line options to control execution, we need to test the validity of each command-line argument and then parse through each one to set up how the script is to be executed. We have four tests that need to be performed to validate each argument. Validating the Number of Command-Line Arguments The first step is to ensure that the number of command-line arguments is what we are expecting. For this script we are expecting no more than three arguments. To test the number of arguments, we use the echo $# command to display the number of command-line arguments. The result is greater than or equal to 0, zero. This test code is shown here.

Free & Share & Open

276

Chapter 10
# Check command line arguments - $# < 3 if (($# > 3)) then usage exit 1 fi

Notice that we used the mathematical test here. One thing to note about the syntax of this test is that for user-, or script-defined variables we do not use the dollar sign ($) in front of the variable. For shell variables you must use the shell notation here, too. If the number of arguments on the command line exceeds three, then we display the usage function and exit the shell script with a return code of 1, one. Test for Valid Command-Line Arguments We really have only three valid command-line arguments. Because -n and -m are lowercase alphabetic characters, we may as well add their uppercase counterparts for people who love to type uppercase characters. Now we have only five valid commandline arguments:
■ ■ ■ ■ ■ ■

Any Integer -n and -N to indicate creating a new $KEYBOARD_FILE -m and -M to indicate that the manager’s password report is to be printed

This seems easy enough to test for using a case statement to parse through the command-line arguments using the $@ values, which is a list of the command-line arguments separated by a single space. Look at the block of code in Listing 10.7 for details.
# Test for valid command line arguments # Valid auguments are “-n, -N, -m, -M, and any integer if (($# != 0)) then for CMD_ARG in $@ do case $CMD_ARG in +([-0-9])) # The ‘+([-0-9]))’ test notation is looking for # an integer. Any integer is assigned to the # length of password variable, LENGTH LENGTH=$CMD_ARG ;; : # The colon (:) is a no-op, which does nothing ;;

-n|-N)

Listing 10.7 Code for testing for command-line arguments.

Creating Pseudo-Random Passwords

277

-m|-M) ;; *) exit

:

# The colon (:) is a no-op, which does nothing # Invalid Command-Line Argument, show usage and

usage exit 1 ;; esac done fi

Listing 10.7 Code for testing for command-line arguments. (continued)

Before we test the validity of each argument, we ensure that there is at least one command-line argument to test. If we have some arguments to test, we start a case statement to parse through each argument on the command line. As the arguments are parsed, the value is assigned to the CMD_ARG variable. Notice the very first test, +([0-9]). This regular expression is testing for an integer value. When we add this integer test to the case statement, we need to add the last close parentheses , ), for the case statement. If the test is true, we know that an integer has been supplied that overrides the default eight-character password length, specified by the LENGTH variable. The tests for -n, -N, -m, and -M are do nothings, or no-ops in this case. A no-op is specified by the colon character (:). The no-op does not do anything, but it always has a 0, zero, return code. When our valid command options are found, the case statement goes to the next argument on the command line. When an invalid command-line option is detected, the function displays the usage message and exits the script with a return code of 1, one, which is defined as a usage error. Ensuring the $LENGTH Variable Is an Integer As a final sanity check of the $LENGTH variable, I added this extra step to ensure that it is assigned an integer value. This test is similar to the test in the previous section, but it is restricted to testing the LENGTH variable assignment. This test code is shown in Listing 10.8.
# # Ensure that the $LENGTH variable is an integer # case $LENGTH in +([0-9])) : # The ‘+([0-9]))’ test notation is looking for # an integer. If it is an integer then the

Listing 10.8 Testing $LENGTH for an integer value. (continues)

Free & Share & Open

278

Chapter 10

# # # # ;; *) usage exit 1 ;; esac

no-op, specified by a colon, (Do Nothing) command is executed, otherwise this script exits with a return code of 1, one, after displaying the usage message

Listing 10.8 Testing $LENGTH for an integer value. (continued)

If the LENGTH variable does not have an integer assignment, then the usage message function is shown, and the script exits with a return code of 1, which is defined as a usage error. Parsing Command-Line Arguments with getopts The getopts function is the best tool for parsing through command-line arguments. With the getopts function we can take direct action or set variables as a valid command-line arguments is found. We can also find invalid command-line arguments, if they are preceded with a minus sign (-). The getopts function is used with a while loop that contains a case statement. The basic syntax is shown in Listing 10.9. while getopts “:n N V: m M” AUGEMENT 2>/dev/null 2>&1 do case $ARGUMENT in n|N) # Do stuff for -n and -N ;; m|M) # Do stuff for -m and -M ;; V) # The colon (:) after the V, V:, specifies # that -V must have an option attached on the command line. ;; \?) # The very first colon (:n) specifies that any unknown # argument (-A, for example) produces a question mark (?) as # output. For these unknown arguments we show the usage # message and exit with a return code of 1, one. ;; esac done

Listing 10.9 Basic syntax for using the getopts function.

Creating Pseudo-Random Passwords
As you can see, using getopts to parse command-line arguments is an easy way to catch invalid command-line arguments and also to assign values or tasks to specific arguments. The nice thing about this method is that we do not have to worry about the order of the arguments on the command line. Let’s look at the code for parsing the command line for this shell script, as shown in Listing 10.10.
# Use the getopts function to parse the command# line arguments. while getopts “:n N m M” ARGUMENT 2>/dev/null do case $ARGUMENT in n|N) # Create a new Keyboard Data file load_default_keyboard echo “\nPress ENTER when you are you ready to continue: \c” read REPLY clear ;; m|M) # Print the Manager Password Report PRINT_PASSWORD_MANAGER_REPORT=TRUE ;; \?) # Show the usage message usage exit 1 esac done

279

Listing 10.10 getops command line parsing.

In our getopts statement, located on the line with the while loop, notice that there is only one colon (:) in the listing. This specifies that any invalid option is to be assigned the question mark (?), specifying an unknown option. We do not have any colons after any options so we are not expecting any values to be assigned to any arguments. In the case of the -n and -N options the load_default_keyboard function is executed. For the -m and -M options the printer variable is set to TRUE. Any other options result in the script exiting with a return code of 1, one.

Beginning of Main
Now that we have defined all of the variables and functions and verified all of the command-line arguments, we are ready to start the main part of the mk_passwd.ksh shell script.

Free & Share & Open

280

Chapter 10

Setting a Trap
The first thing to do is to set a trap. A trap allows us to take action before the shell script or function exits, if an exit signal is trappable and defined. We can never trap a kill -9 exit. This kill option does not do anything graceful; it just removes the process from the system process table, and it no longer exists. The more common exit signals are 1, 2, 3, and 15. For a complete list of exit signals see Chapter 1, or enter kill -l (that’s ell) on the command line. Our trap is shown here: trap ‘trap_exit; exit 2’ 1 2 3 15

When a trapped exit signal is detected, in this case signals 1, 2, 3, or 15, the trap executes the two commands enclosed within the single tic marks, (‘ commands ‘). The commands include running the trap_exit function that removes the $OUTFILE file; then the script exits with a return code of 2, which has been defined as a trap exit for this shell script.

Checking for the Keyboard File
This shell script is useless without a keyboard data file and cannot execute anything. To check for the existence of the $KEYBOARD_FILE, we execute the check_for_ and_create_keyboard_file function. As we previously saw, this function checks to see if a keyboard data file is on the system. If the file is not found, then the user is prompted to automatically load the default keyboard layout, which is a standard 109 key QWERT keyboard. This functionality allows for a quick start for new users and an easy recovery if the file is deleted. When we want to load a custom keyboard layout, all that is needed is to replace the default keyboard file with a new keyboard layout file.

Loading the “KEYS” Array
Once we have a $KEYBOARD_FILE we are ready to load the KEYS array with the keyboard characters. For this shell script we are loading the KEYS array with file data. The easiest way to do this is to use a while loop to read each line of the file, which in this case is a single character, while feeding the loop from the bottom, as shown in Listing 10.11.
X=0 # Initialize the array counter to zero # Load the array called “KEYS” with keyboard elements # located in the $KEYBOARD_FILE. while read ARRAY_ELEMENT do

Listing 10.11 Code to load the KEYS array.

Creating Pseudo-Random Passwords

281

((X = X + 1)) # Increment the counter by 1 # Load an array element in the array KEYS[$X]=$ARRAY_ELEMENT done < $KEYBOARD_FILE

UPPER_LIMIT=$X

# Random Number Upper Limit

Listing 10.11 Code to load the KEYS array. (continued)

In Listing 10.11 we initialize a loop counter, X, to zero. This counter is used to index each array element in sequential order. Next we start the while loop to read each line of data, a single character, and assign the value to the ARRAY_ELEMENT variable on each loop iteration. Inside of the while loop the counter is incremented as the loop progresses, and the KEYS array is assigned a new array element on each loop iteration until all of the file data is loaded into the KEYS array. Notice the command syntax we use to load an array element.
KEYS[$X]=$ARRAY_ELEMENT

At the bottom of the while loop after done, notice the input redirection into the loop. This is one of the fastest ways to parse a file line by line. For more information on this and other file parsing methods, see Chapter 2. The last task is to define the UPPER_LIMIT variable. This variable is used to create the pseudo-random numbers that are used to point to the KEYS array elements when creating a new pseudo-random password.

Using the LENGTH Variable to Build a Loop List
A for loop needs a list of something to loop through, which is defined on the for loop declaration line. This next section of code uses the $LENGTH value to create a list of numbers to loop through. This list of numbers represents the length of the password. The default list is 1 2 3 4 5 6 7 8. The code to build this list is shown in Listing 10.12.
# Produce the “for” loop list of elements that represent # the length of the password: ‘1 2 3 4 5 6 7 8’ is # the default “for” loop list. FOR_COUNT=$(

Listing 10.12 Code to build a for loop list of numbers. (continues)

Free & Share & Open

282

Chapter 10

X=0 while ((X < LENGTH)) do # Build the list here ((X = X + 1)) echo “$X “ done )

Listing 10.12 Code to build a for loop list of numbers. (continued)

Notice how the command substitution is used in Listing 10.12. The entire while loop is enclosed within a command substitution, specified by the MY_LIST=$( all of my commands ) syntax. The while loop is interesting. This is a good way to build a list. The process consists of incrementing a counter and then using an echo or print command to print the character, followed by a blank space. The result is a list of characters separated by a single space.

Building a New Pseudo-Random Password
The code to build a new password is short and relatively easy to understand. The code is shown in Listing 10.13. After the code listing, we will cover the details.
# Create the pseudo-random password in this section clear PW= # Clear the screen # Initialize the password to NULL

# Build the password using random numbers to grab array # elements from the KEYS array. for i in $FOR_COUNT do PW=${PW}${KEYS[$(in_range_random_number $UPPER_LIMIT)]} done # Done building the password

Listing 10.13 Building a new pseudo-random password code.

Creating Pseudo-Random Passwords
We first initialize the password variable (PW) to a null value, specified by PW= , when you make a variable assign to nothing, then you set the variable to NULL. Next we use a for loop to loop through the numbers we previously created and assigned to the FOR_COUNT variable. The default value for this variable is 1 2 3 4 5 6 7 8. Inside the for loop we use a single command to build the password by adding a new pseudo-random character as we go through each loop iteration. Building the password works like this. We start with a NULL variable, PW. On each loop iteration we assign the PW variable the previous PW assignment, which it had from the last loop iteration. Then we add to this current character string a new character, which we generate using the in_range_random_number function inside the KEYS array element assignment using command substitution. The in_range_random_number function expects as input the $UPPER_LIMIT value, which is 109 keys for the default keyboard layout in this script. Using this method we use the function directly in the KEY array element assignment. This is a good way to build a list.

283

Printing the Manager’s Password Report for Safe Keeping
This last section of code will create a temporary report file for printing purposes. The only time this section of code is executed is when the -m or -M command-line arguments are present. In the getops command-line parsing section, the PRINT_PASSWORD_ MANAGER_REPORT variable is assigned the value TRUE. Any other value disables the printing option. This section of code, shown in Listing 10.14, tests the printing variable and if TRUE, executed the build_manager_password_report function. The user is then prompted to print to the default printer, which is listed in the text. The user has a chance to change the printer/queue at this point or to cancel the printing completely. If the $OUTFILE is printed, the lp command adds the -c switch to make a copy of the file in the spooler. This method allows us to immediately delete the password report file from the system. We just do not want this report file sitting on the system for very long.

# Print the Manager’s password report, if specified # on the command with the -m command switch. if [ $PRINT_PASSWORD_MANAGER_REPORT = TRUE ] then typeset -u REPLY=N echo “\nPrint Password Sheet for the Secure Envelope? (Y/N)? \c”

Listing 10.14 Code to create and print the password report. (continues)

Free & Share & Open

284

Chapter 10

read REPLY if [[ $REPLY = ‘Y’ ]] then build_manager_password_report REPLY= # Set REPLY to NULL

echo “\nPrint to the Default Printer ${DEFAULT_PRINTER} (Y/N)? \c” read REPLY if [[ $REPLY = ‘Y’ ]] then echo “\nPrinting to $DEFAULT_PRINTER\n” lp -c -d $DEFAULT_PRINTER $OUTFILE else echo “\nNEW PRINT QUEUE: \c” read DEFAULT_PRINTER echo “\nPrinting to $DEFAULT_PRINTER\n” lp -c -d $DEFAULT_PRINTER $OUTFILE fi else echo “\n\n\tO.K. - Printing Skipped...” fi fi #################################################### # # Remove the $OUTFILE, if it exists and has a size # greater than zero bytes. [ -s $OUTFILE ] && rm -f $OUTFILE

Listing 10.14 Code to create and print the password report. (continued)

The last two things that are done at the end of this shell script are to remove the $OUTFILE, if it exists, and then prompt the user to press ENTER to clear the screen and exit. We do not want to leave a password on the screen for anyone to read. That is it for the steps involved to create the mk_passwd.ksh shell script. The entire shell script is shown in Listing 10.15. Pay particular attention to the boldface text throughout the mk_passwd.ksh shell script.
#!/usr/bin/ksh # # AUTHOR: Randy Micahel # SCRIPT: mk_passwd.ksh

Listing 10.15 mk_passwd.ksh shell script listing.

Creating Pseudo-Random Passwords

285

# DATE: 11/12/2001 # REV: 1.2.P # # PLATFORM: Not Platform Dependent # # PURPOSE: This script is used to create pseudo-random passwords. # An external keyboard data file is utilized, which is # defined by the KEYBOARD_FILE variable. This keyboard # file is expected to have one character on each line. # These characters are loaded into an array, and using # pseudo-random numbers generated, the characters are # “randomly” put together to form a string of characters. # By default, this script produces eight-character passwords, # but this length can be changed on the command line by # adding an integer value after the script name. There are # two command-line options, -n, which creates the default # KEYBOARD_FILE, and -m, which prints the manager’s # password report. This password report is intended # to be locked in a safe for safe keeping. # # EXIT CODES: # 0 - Normal script execution # 1 - Usage error # 2 - Trap exit # 3 - Missing Keyboard data file # # REV LIST: # 6/26/2002: Added two command-line options, -n, which # creates a new $KEYBOARD_FILE, and -m, which prints # the manager’s password report. # # set -x # Uncomment to debug # set -n # Uncomment to check syntax without any command execution # #################################################### ########### DEFINE SOME VARIABLES HERE ############# #################################################### LENGTH=8 # Default Password Length # Notification List for Printing the Manager’s # Password Report for Locking Away Passwords # Just in Case You Are Unavaliable. NOTIFICATION_LIST=”Donald Duck, Yogi Bear, and Mr. Ranger” # Define the Default Printer for Printing the Manager’s # Password Report. The user has a chance to change this

Listing 10.15 mk_passwd.ksh shell script listing. (continues)

Free & Share & Open

286

Chapter 10

# printer at execution time. DEFAULT_PRINTER=”hp4@yogi” SCRIPT=$(basename $0) OUTFILE=”/tmp/tmppdw.file” KEYBOARD_FILE=/scripts/keyboard.keys PRINT_PASSWORD_MANAGER_REPORT=”TO_BE_SET” RANDOM=$$ # Initialize the random number seed to the # process ID (PID) of this shell script.

#################################################### ########## DEFINE FUNCTIONS HERE ################### #################################################### function in_range_random_number { # Create a pseudo-random number less than or equal # to the $UPPER_LIMIT value, which is defined in the # main body of the shell script. RANDOM_NUMBER=$(($RANDOM % $UPPER_LIMIT + 1)) echo “$RANDOM_NUMBER” } # #################################################### # function load_default_keyboard { # If a keyboard data file does not exist then the user # prompted to load the standard keyboard data into the # $KEYBOARD_FILE, which is defined in the main body of # the shell script. clear # Clear the screen

echo “\nLoad the default keyboard data file? (Y/N): \c” read REPLY case $REPLY in y|Y) : ;; *) echo “\nSkipping the load of the default keyboard file...\n”

Listing 10.15 mk_passwd.ksh shell script listing. (continued)

Creating Pseudo-Random Passwords

287

return ;; esac cat /dev/null > $KEYBOARD_FILE echo “\nLoading the Standard Keyboard File...\c” # Loop through each character in the following list and # append each character to the $KEYBOARD_FILE file. This # produces a file with one character on each line. for CHAR in \` 1 2 3 4 5 6 7 8 9 0 - = \\ q w e r t y u i o \ p \[ \] a s d f g h j k l \; \’ z x c v b n m \, \ \. \/ \\ \~ \! \@ \# \$ \% \^ \& \* \( \) _ \+ \| \ Q W E R T Y U I O P \{ \} A S D F G H J K L \: \” \ Z X C V B N M \< \> \? \| \. 0 1 2 3 4 5 6 7 8 9 \/ \ \* \- \+ do echo “$CHAR” >> $KEYBOARD_FILE done echo “\n\n\t...Done...\n” sleep 1 } # #################################################### # function check_for_and_create_keyboard_file { # If the $KEYBOARD_FILE does not exist then # ask the user to load the “standard” keyboard # layout, which is done with the load_default_keyboard # function. if [ ! -s $KEYBOARD_FILE ] then echo “\n\nERROR: Missing Keyboard File” echo “\n\nWould You Like to Load the” echo “Default Keyboard Layout?” echo “\n\t(Y/N): \c” typeset -u REPLY=FALSE read REPLY if [ $REPLY != Y ] then echo “\n\nERROR: This shell script cannot operate” echo “without a keyboard data file located in” echo “\n==> $KEYBOARD_FILE\n”

Listing 10.15 mk_passwd.ksh shell script listing. (continues)

Free & Share & Open

288

Chapter 10

echo “\nThis file expects one character per line.” echo “\n\t...EXITING...\n” exit 3 else load_default_keyboard echo “\nPress ENTER when you are you ready to continue: \c” read REPLY clear fi fi } # #################################################### # function build_manager_password_report { # Build a file to print for the secure envelope ( echo “\n RESTRICTED USE!!!” echo “\n\n\tImmediately send an e-mail to:\n” echo “ $NOTIFICATION_LIST”

echo “\n\tif this password is revealed!” echo “\n\tAIX root password: $PW\n” echo “\n\n” echo “\n RESTRICTED USE!!!” echo “\n\n\tImmediately send an e-mail to:\n” echo “ $NOTIFICATION_LIST”

echo “\n\tif this password is revealed!” echo “\n\tAIX root password: $PW\n” echo “\n\n” echo “\n RESTRICTED USE!!!” echo “\n\n\tImmediately send an e-mail to:\n” echo “ $NOTIFICATION_LIST”

echo “\n\tif this password is revealed!”

Listing 10.15 mk_passwd.ksh shell script listing. (continued)

Creating Pseudo-Random Passwords

289

echo “\n\tAIX root password: ) > $OUTFILE

$PW\n”

} # #################################################### # function usage { echo “\nUSAGE: $SCRIPT [-m] [-n] [password_length]\n” echo “ Where: -m -n Creates a password printout for Security Loads the default keyboard data keys file

password_length - Interger value that overrides the default 8 character password length.\n” } # #################################################### # function trap_exit { rm -f $OUTFILE >/dev/null 2>&1 } #################################################### ########## END OF FUNCTION DEFINITIONS ############# #################################################### #################################################### ####### VALIDATE EACH COMMAND LINE ARGUMENT ######## #################################################### # Check command line arguments - $# < 3 if (($# > 3)) then usage exit 1 fi

Listing 10.15 mk_passwd.ksh shell script listing. (continues)

Free & Share & Open

290

Chapter 10

#################################################### # # Test for valid command line arguments # Valid auguments are “-n, -N, -m, -M, and any integer if (($# != 0)) then for CMD_ARG in $@ do case $CMD_ARG in +([-0-9])) # The ‘+([-0-9]))’ test notation is looking for # an integer. Any integer is assigned to the # length of password variable, LENGTH LENGTH=$CMD_ARG ;; -n) : ;; -N) -m) -M) *) usage exit 1 ;; esac done fi #################################################### # # Ensure that the $LENGTH variable is an integer case $LENGTH in +([0-9])) : # The ‘+([-0]))’ test notation is looking for # an integer. If an integer then the # no-op, specified by a colon, (Do Nothing) # command is executed, otherwise this script # exits with a return code of 1, one. ;; *) usage : ;; : ;; : ;;

Listing 10.15 mk_passwd.ksh shell script listing. (continued)

Creating Pseudo-Random Passwords

291

exit 1 ;; esac #################################################### # # Use the getopts function to parse the command# line arguments. while getopts “:n N m M” ARGUMENT 2>/dev/null do case $ARGUMENT in n|N) # Create a new Keyboard Data file load_default_keyboard echo “\nPress ENTER when you are you ready to continue: \c” read REPLY clear ;; m|M) # Print the Manager Password Report PRINT_PASSWORD_MANAGER_REPORT=TRUE ;; \?) # Show the usage message usage exit 1 esac done #################################################### ################ START OF MAIN ##################### #################################################### # Set a trap trap ‘trap_exit;exit 2’ 1 2 3 15 #################################################### # # Check for a keyboard data file check_for_and_create_keyboard_file #################################################### ############### LOAD THE ARRAY #####################

Listing 10.15 mk_passwd.ksh shell script listing. (continues)

Free & Share & Open

292

Chapter 10

#################################################### X=0 # Initialize the array counter to zero # Load the array called “KEYS” with keyboard elements # located in the $KEYBOARD_FILE. while read ARRAY_ELEMENT do ((X = X + 1)) # Increment the counter by 1 # Load an array element in the the array KEYS[$X]=$ARRAY_ELEMENT done < $KEYBOARD_FILE

UPPER_LIMIT=$X

# Random Number Upper Limit

#################################################### # # Produce the “for” loop list of elements that represent # the length of the password: ‘1 2 3 4 5 6 7 8’ is # the default “for” loop list. FOR_COUNT=$( X=0 while ((X < LENGTH)) do # Build the list here ((X = X + 1)) echo “$X “ done ) #################################################### # # Create the pseudo-random password in this section clear PW= # Clear the screen # Initialize the password to NULL

# Build the password using random numbers to grab array

Listing 10.15 mk_passwd.ksh shell script listing. (continued)

Creating Pseudo-Random Passwords

293

# elements from the KEYS array. for i in $FOR_COUNT do PW=${PW}${KEYS[$(in_range_random_number $UPPER_LIMIT)]} done # Done building the password #################################################### # # Display the new pseudo-random password to the screen echo “\n\n echo “\n The new $LENGTH character password is:\n” ${PW}\n”

#################################################### # # Print the Manager’s password report, if specified # on the command with the -m command switch. if [ $PRINT_PASSWORD_MANAGER_REPORT = TRUE ] then typeset -u REPLY=N echo “\nPrint Password Sheet for the Secure Envelope? (Y/N)? \c” read REPLY if [[ $REPLY = ‘Y’ ]] then build_manager_password_report REPLY= # Set REPLY to NULL

echo “\nPrint to the Default Printer ${DEFAULT_PRINTER} (Y/N)? \c” read REPLY if [[ $REPLY = ‘Y’ ]] then echo “\nPrinting to $DEFAULT_PRINTER\n” lp -c -d $DEFAULT_PRINTER $OUTFILE else echo “\nNEW PRINT QUEUE: \c” read DEFAULT_PRINTER echo “\nPrinting to $DEFAULT_PRINTER\n”

Listing 10.15 mk_passwd.ksh shell script listing. (continues)

Free & Share & Open

294

Chapter 10

lp -c -d $DEFAULT_PRINTER $OUTFILE fi else echo “\n\n\tO.K. - Printing Skipped...” fi fi #################################################### # # Remove the $OUTFILE, if it exists and has a size # greater than zero bytes. [ -s $OUTFILE ] && rm -f $OUTFILE #################################################### # # Clear the screen and exit echo “\n\nPress ENTER to Clear the Screen and EXIT: \c” read X clear # End of mk_passwd.ksh shell script

Listing 10.15 mk_passwd.ksh shell script listing. (continued)

This was an interesting shell script to create. I hope you picked up some pointers in this chapter. I tried to add as many script options to this script as desirable but not make the script too difficult to understand.

Other Options to Consider
As with any script, improvements can be made. I cannot think of anything to add to the script, but you may want to remove some of the functionality for the common user community.

Password Reports?
Do you need to create password reports for your Manager and Directors? If not, you should disable the ability to create any file that contains any password and disable printing any passwords. This is easy to disable by commenting out the getopts parsing for the -m and -M command-line options.

Creating Pseudo-Random Passwords

295

Which Password?
You certainly do not have to accept the first password that is produced by this script. It usually takes me 5 to 10 tries to get a password that I may be able to remember. Don’t stop at the first one—keep going until you get a password that you like but is not guessable.

Other Uses?
Sure, there are other uses for this shell script. Any time that you need a pseudo-random list of keyboard characters, you can use this shell script to create the list. License key is the first thing that comes to mind. If you are selling software and you need to create some unguessable keys, run the script and specify the length of the key as an integer value.

Summary
This was an excellent exercise in creating pseudo-random numbers and using a function directly in a command assignment. We used arrays to store our keyboard data so that any element is directly accessible. This chapter goes a long way in making any task intuitively obvious to solve. We love a good challenge. In the next chapter we are going to study how to monitor for stale disk partitions on an AIX system. I’ll see you in the next chapter!

Free & Share & Open

CHAPTER

11
Monitor for Stale Disk Partitions

Monitoring for stale disk partitions is an AIX thing. To understand this chapter you need to be familiar with the Logical Volume Manager (LVM) that is at the heart of the AIX operating system. We will get to the LVM in the next section. At the high level a stale disk partition means that the mirrored disks are not in sync. Sometimes when you find stale disks partitions you can resync the mirrors, and all is well. If the mirrors will not sync up, you may be seeing the first signs of a failing disk. In this chapter we are going to look at three methods of monitoring for stale partitions:
■■ ■■ ■■

Monitoring at the Logical Volume (LV) level Monitoring at the Physical Volume (PV), or disk, level Monitoring at the Volume Group (VG), PV, and LV levels to get the full picture

All three methods will report the number of stale disk partitions, but it is nice to know the VG, PV, and the LV that are involved in the unsynced mirrors. We are going to step through the entire process of building these shell scripts, starting with the command syntax required to query the system. Before we start our scripting effort, I want to give you a high-level overview of the AIX LVM and the commands we are going to use.

297

Free & Share & Open

298

Chapter 11

AIX Logical Volume Manager (LVM)
Unlike most Unix operating systems, IBM manages disk resources using a program called the Logical Volume Manager (LVM). The LVM consists of the following components, starting with the smallest. Each Physical Volume (PV), or disk, in the system is broken down into small partitions called Physical Partitions (PP). The default size of a PP is 4MB, but it can be larger depending on the size of the disk. The LVM uses groups of these PPs to create a logical map to point to the actual PPs on the disk. These mapped partitions are called Logical Partitions (LP). The sizes of an LP and PP are exactly the same because an LP is just a pointer to a PP. At the next level we have the Logical Volume (LV). An LV consists of one or more LPs. The LV can span multiple PVs, and this is what differentiates AIX from other flavors of Unix. This is the level at which the Systems Administrator creates the mirrors. When an LV is first created, the LV is considered raw, meaning that it does not have a Filesystem mount point. Raw LVs are commonly used for databases. On top of an LV we can create a Filesystem, which has a mount point—for example, /scripts. The LV does not require a Filesystem if you want the LV to remain raw, but you can create one. Volume Group (VG) is a collection of one or more Physical Volumes (PV), or disks. A PV is listed on the system as an hdisk#, where # is an integer value. A VG is the largest component of the LMV. The VG contains one or more LVs, so this is the mechanism that allows an LV to span multiple PVs. That is the high-level overview of the LVM and its components. For this chapter we are going to focus our attention at the VG, PV, LV, and PP levels, and we are concerned only with disks in a mirrored configuration. If you want more information on the AIX LVM there are plenty of books that go into great detail about AIX system management.

The Commands and Methods
As usual, we need the command syntax before we can write a shell script. We will work with three LVM commands in this chapter. Each of these commands queries the system for specific information on the components and status of the disk subsystem. Before we proceed, it is important to know what each of these commands is used for and what type of information can be gathered from the system.

Disk Subsystem Commands
The lsvg command queries the system for VG information. To see which VGs are varied-on, or active, we add the -o switch to the lsvg command. We also have the -l flag that allows the lsvg command to query the system for the contents of a specific VG. We are interested in one of the fields in the lsvg command output called STALE_PPs:, which has a value representing the number of stale PVs in the target VG. Ideally we want this number to be zero.

Monitor for Stale Disk Partitions
Then we move to the LV command, lslv. The lslv command will query the system for the status information of a specific LV, which is entered as a command parameter. One of the fields in the output of the lslv command is STALE PP:. This output shows the number of stale PPs for the LV specified on the command line. Ideally, we want this number to be 0, zero. If we add the -l flag to the lslv command, we can see which PVs are associated with the LV in the first column of the command output. Next we can move down to the PV, or disk, level. The lspv command queries the system for information on a specific PV, which is passed as a command parameter to the lspv command. Like lslv, the lspv command also reports the number of STALE PARTITIONS: as a field in the output. You will see the output of each of these commands as we write the scripts for this chapter. We have the commands defined so we are now ready to start creating our first shell script to monitor for stale disk partitions.

299

Method 1: Monitoring for Stale PPs at the LV Level
The easiest, but not always the quickest, method of checking for stale disk partitions is to work at the LV level of the LVM structure. Querying the system for LV stale partition information gives the high-level overview for each LV. If however, the LV spans more than one PV, or disk, then another step must be taken to find the actual mirrored disks that are not in sync. We will get to this finer granularity of monitoring in the next section of this chapter. We start our monitoring by issuing an LVM query to find each of the active VGs on the system, or the VGs that are varied online. For this step we use the lsvg -o command. The -o flag tells the lsvg command to list only the volume groups that are currently varied online. Many more VGs may exist on the system, but if they are not varied online we cannot check the status of any of the logical volumes that reside within the VG because the entire VG is inactive. Let’s assign the VG list to a variable called ACTIVE_VG_LIST.
ACTIVE_VG_LIST=$(lsvg -o)

My test machine has two VGs, and both are active: rootvg appvg2

The previous command saves the active Volume Groups in a variable. Using the ACTIVE_VG_LIST variable contents we next create a list of active LVs on the system. Each VG will have one or more LVs that may or may not be active, or open. Using the $ACTIVE_VG_LIST data we can query the system to list each active LV within each active VG. The lsvg -l $VG command queries the system at the VG level to display the contents. Listing 11.1 shows the output of lsvg -l appvg2 rootvg command on my test machine.

Free & Share & Open

300

Chapter 11

appvg2: LV NAME tel_lv oracle_lv oradata_lv ar_lv remp_tmp01 export_lv loglv00 remp2_ctl01 remp2_ctl02 remp2_ctl03 rempR2_dat01 R2remedy_lv remp2_log1a remp2_log1b remp2_log2a remp2_log2b remp2_log3a remp2_log3b remp2_log4a remp2_log4b remp2_log5a remp2_log5b remp2_rbs01 remp2_sys01 arlogs_lv remp2_usr01 rootvg: LV NAME hd5 hd6 hd8 hd4 hd2 hd9var hd3 hd1 local_lv

TYPE jfs jfs jfs jfs jfs jfs jfslog jfs jfs jfs jfs jfs jfs jfs jfs jfs jfs jfs jfs jfs jfs jfs jfs jfs jfs jfs TYPE boot paging jfslog jfs jfs jfs jfs jfs jfs

LPs 2 128 128 16 128 100 1 1 1 1 192 10 1 1 1 1 1 1 1 1 1 1 47 4 35 6 LPs 1 80 1 4 40 10 10 3 9

PPs 2 128 128 16 128 100 1 1 1 1 192 10 1 1 1 1 1 1 1 1 1 1 47 4 35 6 PPs 2 160 2 8 80 20 20 6 18

PVs 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 PVs 2 2 2 2 2 2 2 2 2

LV STATE open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd LV STATE closed/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd

MOUNT POINT /usr/telalert /oracle /oradata /usr/ar /remd_tmp01 /export N/A /remd_ctl01 /remd_ctl02 /remd_ctl03 /remd_dat01 /usr/remedy /remd_log1a /remd_log1b /remd_log2a /remd_log2b /remd_log3a /remd_log3b /remd_log4a /remd_log4b /remd_log5a /remd_log5b /remd_rbs01 /remd_sys01 /usr/ar/logs /remd_usr01 MOUNT POINT N/A N/A N/A / /usr /var /tmp /home /usr/local

Listing 11.1 Output of the lsvg -l appvg2 rootvg command.

The list of LVs is shown in column one. Notice the sixth column in the output in Listing 11.1, LV STATE. Most of the LVs are open/synced, but one LV, hd5, is closed/ synced. The hd5 LV that is closed is the boot logical volume and is active only when the system is booting up. Because we want only active LVs all we need to do is to grep on

Monitor for Stale Disk Partitions the string open and then awk out the first column. The next command saves the list of currently active LVs in a variable called ACTIVE_LV_LIST.
ACTIVE_LV_LIST=$(lsvg -l $ACTIVE_VG_LIST | grep open | awk ‘{print $1}’)

301

In the previous command, we use our $ACTIVE_VG_LIST as a command parameter for the lsvg -l command. Then we pipe ( | ) to grep the lsvg output for only the rows that contain the string open. Next, another pipe is used to awk out the first column, specified by awk '{print $1}'. The result is a list of currently active LV names. If you think about an array, the grep command works on the rows and the awk command works on the columns. The only thing left to do is to query each LV for the number of stale PPs, specified by the STALE PP: field. To check every LV we need to set up a for loop to run the same command on each LV in the active list. The command we use to query the LV is lslv -L $LV_NAME. The output for a single LV is shown in Listing 11.2.
LOGICAL VOLUME: remp_tmp01 LV IDENTIFIER: 00011151b819f83a.5 VG STATE: active/complete TYPE: jfs MAX LPs: 512 megabyte(s) COPIES: 1 LPs: 128 STALE PPs: 0 INTER-POLICY: minimum INTRA-POLICY: middle MOUNT POINT: /remd_tmp01 MIRROR WRITE CONSISTENCY: on EACH LP COPY ON A SEPARATE PV ?: yes VOLUME GROUP: PERMISSION: LV STATE: WRITE VERIFY: PP SIZE: SCHED POLICY: PPs: BB POLICY: RELOCATABLE: UPPER BOUND: LABEL: appvg2 read/write opened/syncd off 32 parallel 128 relocatable yes 32 /remd_tmp01

Listing 11.2 LV statistics for the remp_tmp01 logical volume.

Notice in the command output in Listing 11.2 the ninth row, where the field STALE PP: is listed. The second column of this row contains the number of stale partitions in the logical volume. Ideally, we want this value to be zero, 0. If the value is greater than zero we have a problem. Specifically, the mirrored disks associated with this LV are not in sync, which translates to a worthless mirror. Looking at this output, how are we supposed to get the number of stale disk partitions? It turns out that this is a very simple combination of grep and awk. Take a look at the following command statement.
NUM_STALE_PP=$(lslv -L $LV | grep “STALE PP” | awk ‘{print $3}’

The previous statement saves the number of stale PPs into the NUM_STALE_PP variable. We accomplish this feat by command substitution, specified by the VARIABLE=$( commands ) notation. The way to make this task easy is to do the

Free & Share & Open

302

Chapter 11 parsing one step at a time. First, the row containing the STALE PP string is extracted and is provided as input to the next command in the pipe. The next command in the pipe is an awk statement that extracts only the third field, specified by '{print $3}'. At this point you may be asking why we used the third field instead of the second. By default, awk uses white space as a field separator, and because STALE PPs: 0 contains two areas of white space, we need the third field instead of the second. Now that we have all of the commands, all we need to do is set up a loop to run the previous command against each logical volume stored in the $ACTIVE_LV_LIST variable. A little for loop will work just fine for this script. The loop is shown in Listing 11.3.
THIS_HOST=$(hostname) for LV in $(echo $ACTIVE_LV_LIST) do NUM_STALE_PP=$(lslv -L $LV | grep “STALE PP” | awk ‘{print $3}’ if ((NUM_STALE_PP > 0)) then echo “\n${THIS_HOST}: $LV has $NUM_STALE_PP stale PPs” fi done

Listing 11.3 Loop to show the number of stale PPs from each LV.

I want to point out several things in Listing 11.3. First, notice that we save the hostname of the machine in a variable called THIS_HOST. When creating any type of report we need to know which machine we are reporting on. When you have more than 100 machines, things can get a little confusing if you do not have a hostname to go with the report. A for loop needs a list of items to loop through. To get the list of active LVs, we use command substitution to echo the contents of the $ACTIVE_LV_LIST to provide our for loop with a list. Actually, the echo is not necessary, but I wanted to show you a varied approach. The next step is to run the lslv -L command for each LV listed and extract the field that shows the number of stale PPs. For this command we again use command substitution to assign the value to a variable called NUM_STALE_PP. Using this saved value we do a numeric test in the if statement. Notice that we did not add a dollar sign ($) in front of the NUM_STALE_PP variable. Because we used the double parentheses numeric test method, the command assumes that every nonnumeric string is a variable so the dollar sign ($) is not needed; in fact, the test may give an error if the $ was added. If we find that the number of stale PPs is greater than zero, then we use an echo statement to show the hostname of the machine followed by the LV name that has stale partitions and, last, the number of stale partitions that were found. These steps are followed for every active LV in every active VG on the entire system. The full shell script is shown in Listing 11.4.

Monitor for Stale Disk Partitions

303

#!/bin/ksh # # SCRIPT: stale_LV_mon.ksh # # AUTHOR: Randy Michael # DATE: 01/22/2002 # REV: 1.1.P # # PLATFORM: AIX only # # PURPOSE: This shell script is used to query the system # for stale PPs in every active LV within every active # VG. # # REVISION LIST: # # # set -x # Uncomment to debug this script # set -n # Uncomment to check command syntax without any execution THIS_HOST=`hostname` STALE_PP_COUNT=0 # Hostname of this machine # Initialize to zero

# Find all active VGs echo “\nGathering a list of active Volume Groups” ACTIVE_VG_LIST=$(lsvg -o) # Find all active LVs in every active VG. echo “\nCreating a list of all active Logical Volume” ACTIVE_LV_LIST=$(lsvg -l $ACTIVE_VG_LIST | grep open | awk ‘{print $1}’) # Loop through each active LV and query for stale disk partitions echo “\nLooping through each Logical Volume searching for stale PPs” echo “...Please be patient; this may take several minutes to complete...” for LV in $(echo $ACTIVE_LV_LIST) do # Extract the number of STALE PPs for each active LV NUM_STALE_PP=`lslv -L $LV | grep “STALE PP” | awk ‘{print $3}’` # Check for a value greater than zero if ((NUM_STALE_PP > 0)) then # Increment the stale PP counter (( STALE_PP_COUNT = $STALE_PP_COUNT + 1)) # Report on all LVs containing stale disk partitions echo “\n${THIS_HOST}: $LV has $NUM_STALE_PP PPs”

Listing 11.4 stale_LV_mon.ksh shell script listing. (continues)

Free & Share & Open

304

Chapter 11

fi done # Give some feedback if no stale disk partitions were found if ((STALE_PP_COUNT == 0)) then echo “\nNo stale PPs were found in any active LV...EXITING...\n” fi

Listing 11.4 stale_LV_mon.ksh shell script listing. (continued)

Notice in the script in Listing 11.4 that we added notification at each step in the process. As always, we need to let the user know what is going on. Before each command I added an echo statement to show the user how we progress through the shell script. I also added a STALE_PP_COUNT variable to give feedback if no stale PPs were found. Now let’s move on to searching for stale PPs at the PV level instead of the LV level.

Method 2: Monitoring for Stale PPs at the PV Level
Checking for stale disk partitions at the LV level will let you know that one or more LVs have stale PPs. To get a better picture of where the unsynced mirrors reside we need to look at the hdisk level. In this section we are going to change the query point for searching for stale Physical Partitions, or PPs, from the Logical Volume to the Physical Volume, or disk level. The time saving in execution time between these two methods is threefold in favor of working directly with the disks by my measurements. On my test machine, an H-80 RS/6000, the LV query took 40.77 seconds in real time, 0.36 seconds of system time, and 0.02 seconds of user time. Using the PV query method I reduced the execution time to 12.77 seconds in real time and 0.17 seconds of system time, and I had the same 0.02 seconds for user time. To understand the LV and PV configuration I have 18 mirrored disks, which are 9 mirror pairs of 9.1GB disk drives, and a total of 32 LVs. Because an LV query takes longer to execute than a PV query, it is understandable that the PV query won. Depending on the system configuration, this timing advantage may not always hold. In the PV monitoring method we still are concerned only with the hdisks that are in currently varied-on Volume Groups (VGs), as we did in the LV method using the lsvg -o command. Using this active VG list we can query each active VG and extract all of the hdisks that belong to each VG. Once we have a complete list of all of the hdisks we can start a loop and query each of the PVs independently. The output of a PV query is similar to the LV query statistics in Listing 11.2. Take a look at the PV query of hdisk5 using the lspv -l hdisk5 command in Listing 11.5.

Monitor for Stale Disk Partitions

305

PHYSICAL VOLUME: PV IDENTIFIER: PV STATE: STALE PARTITIONS: PP SIZE: TOTAL PPs: FREE PPs: USED PPs: FREE DISTRIBUTION: USED DISTRIBUTION:

hdisk5 VOLUME GROUP: appvg2 00011150e33c3f14 VG IDENTIFIER 00011150e33ce9bb active 0 ALLOCATABLE: yes 16 megabyte(s) LOGICAL VOLUMES: 2 542 (8672 megabytes) VG DESCRIPTORS: 1 397 (6352 megabytes) 145 (2320 megabytes) 89..00..91..108..109 20..108..17..00..00

Listing 11.5 PV statistics for the hdisk5 physical volume.

In the output in Listing 11.5 the STALE PARTITIONS: field in row four and its value are the third field in the row. If the stale partition value ever exceeds zero, then we use the same type of reporting technique that we used in the LV query in Method 1. If no stale partitions are found, then we can give the “all is well” message and exit the script. Because we have the basic idea of the process, let’s take a look at the shell script in Listing 11.6. We will explain the technique in further detail at the end of the code listing.

#!/usr/bin/ksh # # SCRIPT: stale_PP_mon.ksh # # AUTHOR: Randy Michael # DATE: 01/29/02 # REV: 1.2.P # # PLATFORM: AIX only # # PURPOSE: This shell script is used to query the system for stale PPs. # The method queries the system for all of the currently # varied-on volume groups and then builds a list # of the PVs to query. If a PV query detects any stale # partitions notification is sent to the screen. Each step in # the process has user notification # # REVISION LIST: # #

Listing 11.6 stale_PP_mon.ksh shell script listing. (continues)

Free & Share & Open

306

Chapter 11

# set -x # Uncomment to debug this shell script # set -n # Uncomment to check command syntax without any execution THIS_HOST=$(hostname) HDISK_LIST= STALE_PP_COUNT=0 # Hostname of this machine # Initialize to NULL # Initialize to zero

# Inform the user at each step echo “\nGathering a list of hdisks to query\n” # Loop through each currently varied-on VG for VG in $(lsvg -o) do # Build a list of hdisks that belong to currently varied-on VGs echo “Querying $VG for a list of disks” HDISK_LIST=”$HDISK_LIST $(lsvg -p $VG |grep disk \ | awk ‘{print $1}’)” done echo “\nStarting the hdisk query on individual disks\n” # Loop through each of the hdisks found in the previous loop for HDISK in $(echo $HDISK_LIST) do # Query a new hdisk on each loop iteration echo “Querying $HDISK for stale partitions” NUM_STALE_PP=$(lspv -L $HDISK | grep “STALE PARTITIONS:” \ | awk ‘{print $3}’) # Check to see if the stale partition count is greater than zero if ((NUM_STALE_PP > 0)) then # This hdisk has at least one stale partition - Report it! echo “\n${THIS_HOST}: $HDISK has $NUM_STALE_PP Stale Partitions” # Build a list of hdisks that have stale disk partitions STALE_HDISK_LIST=$(echo $STALE_HDISK_LIST; echo $HDISK) fi done # If no stale partitions were found send an “all is good” message ((NUM_STALE_PP > 0)) \ || echo “\n${THIS_HOST}: No Stale PPs have been found...EXITING...\n”

Listing 11.6 stale_PP_mon.ksh shell script listing. (continued)

Monitor for Stale Disk Partitions
We totally changed our viewpoint in our search for stale disk partitions. Instead of working at each LV we are scanning each disk, or PV, independently. The search time on my test machine was three times faster, but my machine configuration does not mean that your system query will be as fast. I want to start at the top of our stale_PP_mon.ksh shell script in Listing 11.6 and work to the bottom. We start off the script by initializing three variables, THIS_HOST (the hostname of the reporting machine), HDISK_LIST (the list of PVs to query, which we initialize to NULL), and STALE_PP_COUNT (the total number of stale disk partitions on all disks, which is initialized to zero). We will show how each of these variables is used as we progress through the script. The next step is to use the list of currently varied-on VGs (using the lsvg -o command) to create a list of currently available hdisks—at least they should be available. We do this in a for loop by appending to the HDISK_LIST variable during each loop iteration. Once we have a list of available system disks, we start a for loop to query each hdisk individually. During the query statement:
NUM_STALE_PP=$(lspv -L $HDISK | grep “STALE PARTITIONS:” | awk ‘{print $3}’) \

307

we capture the number of stale disk partitions by using grep and awk together in the same statement. Just remember that the grep command acts on the rows and the awk statement acts on the columns. On each loop iteration we check the value of the $NUM_STALE_PP variable. If the count is greater than zero we do two things: report the disk to the screen, and append to the STALE_HDISK_LIST variable. Notice how we append to a variable that currently has data in it. By initializing the variable to NULL (specified by VARIABLE= ), by creating an assignment to nothing, we can always append to the variable using the following syntax:
VARIABLE=”$VARIABLE $NEW_VALUE”

Because the $VARIABLE has an initial value of nothing, NULL, then the first value assigned is a new value, and all subsequent values are appended to the VARIABLE variable on each loop iteration. At the end of the script we test the $NUM_STALE_PP variable, which has a running count of all stale disk partitions. If the value is zero, then we let the end user know that everything is OK. Notice how we do the test. We do a numerical test on the $NUM_STALE_PP variable to see if it is greater than zero. If the value is one or more, then the statement is true. On a true statement the logical OR ( || ) passes control to the second part of the statement, which states “No stale PPs have been found.” The logical OR saves an if statement and is faster to execute than an if statement. Now that was a fun little script. We can improve on both scripts that have been presented thus far. There is a procedure to attempt to resync the disks containing stale partitions. In the next section we are going to combine the LV and PP query methods and add in a VG query as the top-level query to search for stale disk partitions. We will also attempt to resync all of the stale LVs that we find, if the ATTEMPT_RESYNC variable is set to TRUE.

Free & Share & Open

308

Chapter 11

Method 3: VG, LV, and PV Monitoring with a resync
We have looked at stale disk mirrors from two angles, but we can look for stale disk partitions at a higher level, the VG level. By using the lsvg command we can find which VG has disks that have stale PPs. Using the lsvg command we can shorten our queries to a limited number of disks, although with Murphy’s Law working, it just might be the largest VG on the planet! The strategy that we want to follow is first to query each active VG for stale PVs, which we find using the lsvg command. Then, for each VG that has the STALE PV: field greater than zero, we query the current VG in the loop to get a list of associated PV, or disks. Using a list of all of the PVs that we found, we conduct a query of each disk to find both the list of LVs the PV is associated with and the value of the STALE PARTITIONS: field. For each PV found to have at least one stale partition, we query the PV for a list of LVs that reside on the current PV. Please don’t get confused now! The steps involved are a natural progression through the food chain to the source. The final result of all of these queries is that we know which VG, PV, and LV have unsynced mirrors, which is the complete picture that we want. The process that we follow in this section is faster to execute and easier to follow, so let’s start. The commands we are going to use are shown in Listing 11.7. lsvg -o lsvg $VG_NAME lsvg -p $VG_NAME lspv $PV_NAME lspv -l $PV_NAME lslv $LV_NAME syncvg $HDISK_LIST syncvg -l $LV_LIST varyonvg Produces a list of active VGs Queries the target VG for status information Produces a list of hdisks that belong to the VG Queries the hdisk specified by $PV_NAME Produces a list of LVs on the target hdisk Queries the target LV for status information Synchronizes the mirrors at the hdisk level Synchronizes the mirrors at the LV level Synchronizes only the stale partitions

Listing 11.7 Command summary for the Method 3 shell script.

Using the nine commands in Listing 11.7 we can produce a fast-executing shell script that produces the full picture of exactly where all of the unsynced mirrors reside, and we can even attempt to fix the problem! For this shell script I want to present you with the entire script; then we will step through and explain the philosophy behind the techniques used. In studying Listing 11.8 pay close attention to the bold text.
#!/usr/bin/ksh # # SCRIPT: stale_VG_PV_LV_PP_mon.ksh # # AUTHOR: Randy Michael

Listing 11.8 stale_VG_PV_LV_PP_mon.ksh shell script listing.

Monitor for Stale Disk Partitions

309

# DATE: 01/29/02 # REV: 1.2.P # # PLATFORM: AIX only # # PURPOSE: This shell script is used to query the system for stale PPs. # The method queries the system for all of the currently # varied-on volume groups and then builds a list of the PVs to # query. If a PV query detects any stale partitions notification # is sent to the screen. Each step in the process has user # notification. # # REVISION LIST: # # # set -x # Uncomment to debug this shell script # set -n # Uncomment to check command syntax without any execution # # EXIT CODES: 0 ==> Normal execution or no stale PP were found # 1 ==> Trap EXIT # 2 ==> Auto resyncing failed # ######################################################################## ######### DEFINE VARIABLES HERE #################### ATTEMPT_RESYNC=FALSE # Flag to enable auto resync, “TRUE” will resync # # # # # # # # # # # Stale PP logfile Hostname of this machine Initialize to zero Initialize to zero Initialize to NULL Initialize to NULL Initialize to NULL Initialize to NULL Initialize to NULL Initialize to NULL Initialize to NULL

LOGFILE=”/tmp/stale_PP_log” THIS_HOST=$(hostname) STALE_PP_COUNT=0 STALE_PV_COUNT=0 HDISK_LIST= INACTIVE_PP_LIST= STALE_PV_LIST= STALE_LV_LIST= STALE_VG_LIST= RESYNC_LV_LIST= PV_LIST=

####################################### #### INITIALIZE THE LOG FILE #### >$LOGFILE # Initialize the log file to empty date >> $LOGFILE # Date the log file was created echo “\n$THIS_HOST \n” >> $LOGFILE # Hostname for this report

#### DEFINE FUNCTIONS HERE ############

Listing 11.8 stale_VG_PV_LV_PP_mon.ksh shell script listing. (continues)

Free & Share & Open

310

Chapter 11

# Trap Exit function function trap_exit { echo “\n\t...EXITING on a TRAPPED signal...\n” } ####################################### # Set a trap... trap ‘trap_exit; exit 1’ 1 2 3 5 15 ####################################### ######### BEGINNING OF MAIN ########### ####################################### # Inform the user at each step # Loop through each currently varied-on VG and query VG for stale PVs. # For any VG that has at least one stale PV we then query the VG # for the list of associated PV and build the $PV_LIST echo “\nSearching each Volume Group for stale Physical Volumes...\c” \ | tee -a $LOGFILE # Search each VG for stale PVs, then build a list of VGs and PVs # that have stale disk partitions for VG in $(lsvg -o) do NUM_STALE_PV=$(lsvg $VG | grep ‘STALE PVs:’ | awk ‘{print $3}’) if ((NUM_STALE_PV > 0)) then STALE_VG_LIST=”$STALE_VG_LIST $VG” PV_LIST=”$PV_LIST $(lsvg -p $VG | tail +3 | awk ‘{print $1}’)” ((STALE_PV_COUNT = STALE_PV_COUNT + 1)) fi done # Test to see if any stale PVs were found, if not then # exit with return code 0 if ((STALE_PV_COUNT == 0))

Listing 11.8 stale_VG_PV_LV_PP_mon.ksh shell script listing. (continued)

Monitor for Stale Disk Partitions

311

then echo $LOGFILE exit else echo PPs...\c” fi # # # # # # Now we have a list of PVs from every VG that reported stale PVs The next step is to query each PV to make sure each PV is in an “active” state and then query each PV for stale PPs. If a PV is found to be inactive then we will not query the PV for stale partitions, but move on to the next PV in the $PV_LIST. “\nNo Stale Disk Mirrors Found...EXITING...\n” | tee -a 0 “\nStale Disk Mirrors Found!...Searching each hdisk for stale \ | tee -a $LOGFILE

for HDISK in $(echo $PV_LIST) do PV_STATE=$(lspv $HDISK | grep ‘PV STATE:’ | awk ‘{print $3}’) if [[ $PV_STATE != ‘active’ ]] then INACTIVE_PV_LIST=”$INACTIVE_PV_LIST $HDISK” fi if ! $(echo $INACTIVE_PV_LIST | grep $HDISK) >/dev/null 2>&1 then NUM_STALE_PP=$(lspv $HDISK | grep ‘STALE PARTITIONS:’ \ | awk ‘{print $3}’) if ((NUM_STALE_PP > 0)) then STALE_PV_LIST=”$STALE_PV_LIST $HDISK” ((STALE_PP_COUNT = $STALE_PP_COUNT + 1)) fi fi done # Now we have the list of PVs that contain the stale PPs. # Next we want to get a list of all of the LVs affected. echo “\nSearching each disk with stale PPs for associated LVs\c” \ | tee -a $LOGFILE for PV in $(echo $STALE_PV_LIST) do STALE_LV_LIST=”$STALE_LV_LIST $(lspv -l $PV | tail +3 \ | awk ‘{print $1}’)”

Listing 11.8 stale_VG_PV_LV_PP_mon.ksh shell script listing. (continues)

Free & Share & Open

312

Chapter 11

done # Using the STALE_LV_LIST variable list we want to query # each LV to find which ones need to be resynced echo “\nSearch each LV for stale partitions to build a resync LV list\c” \ | tee -a $LOGFILE for LV in $(echo $STALE_LV_LIST) do LV_NUM_STALE_PP=$(lslv $LV | grep “STALE PPs:” | awk ‘{print $3}’) (($LV_NUM_STALE_PP == 0)) & RESYNC_LV_LIST=”$RESYNC_LV_LIST $LV” done # If any inactive PVs were found we need to inform the user # of each inactive PV # Check for a NULL variable if [[ -n “$INACTIVE_PV_LIST” && “$INACTIVE_PV_LIST” != ‘’ ]] then for PV in $(echo $INACTIVE_PV_LIST) do echo “\nWARNING: Inactive Physical Volume Found:” | tee -a $LOGFILE echo “\n$PV is currently inactive:\n” | tee -a $LOGFILE echo “\nThis script is not suitable to to correct this problem...” \ | tee -a $LOGFILE echo “ ...CALL IBM SUPPORT ABOUT ${PV}...” | tee -a $LOGFILE done fi echo “\nStale Partitions have been found on at least one disk!” \ | tee -a $LOGFILE echo “\nThe following Volume Group(s) have stale PVs:\n” \ | tee -a $LOGFILE echo $STALE_VG_LIST | tee -a $LOGFILE echo “\nThe stale disk(s) involved include the following:\n” \ | tee -a $LOGFILE echo $STALE_PV_LIST | tee -a $LOGFILE echo “\nThe following Logical Volumes need to be resynced:\n” \ | tee -a $LOGFILE echo $RESYNC_LV_LIST | tee -a $LOGFILE if [[ $ATTEMPT_RESYNC = “TRUE” ]]

Listing 11.8 stale_VG_PV_LV_PP_mon.ksh shell script listing. (continued)

Monitor for Stale Disk Partitions

313

then echo “\nAttempting to resync the LVs on $RESYNC_PV_LIST ...\n” \ | tee -a $LOGFILE syncvg -l $RESYNC_LV_LIST | tee -a $LOGFILE 2>&1 if (( $? == 0)) then echo “\nResyncing all of the LVs SUCCESSFUL...EXITING...” \ | tee -a $LOGFILE else echo “\nResyncing FAILED...EXITING...\n” | tee -a $LOGFILE exit 2 fi else echo “\nAuto resync is not enabled...set to TRUE to automatically resync\n” \ | tee -a $LOGFILE echo “\n\t...EXITING...\n” | tee -a $LOGFILE fi echo “\nThe log file is: $LOGFILE\n”

Listing 11.8 stale_VG_PV_LV_PP_mon.ksh shell script listing. (continued)

The shell script in Listing 11.8 is interesting because of the techniques used. As we start at the top of the shell script, notice the first variable definition, ATTEMPT_ RESYNC. I initialize this variable to FALSE because resyncing at the LV level can cause a significant system load. A better method is to run the varyonvg command without any arguments. This method will only resync the stale partitions. Because of the possibility of loading the system down and slowing production response time, I initialize this variable to FALSE. If I am working on a test/development or sandbox machine, I usually set the ATTEMPT_RESYNC variable to TRUE, in uppercase. The TRUE setting will attempt to resync, at the LV level, of every stale LV. The remaining variables initialize the LOGFILE and THIS_HOST variables to the log filename and hostname, respectively. A couple of counters are initialized to zero, and seven other variables are initialized to NULL. In the next section we initialize the $LOGFILE with header information. The only function in this script is the trap_exit function. The trap_exit function displays only to the screen ...EXITING on a TRAPPED signal... when a trap is captured. The trap is set for exit codes 1, 2, 3, 5, and 15 and then the script exits with return code 1. This functionality is just a notification measure for the user. Now we are at BEGINNING OF MAIN in our script. At each step through this shell script we want to give the user feedback so that he or she will know what is going on. When writing shell scripts you need to do two things: Comment everything and give your users feedback so that they know what is going on. In our first query we inform the users that we are searching each VG for stale PVs. For this step we use the lsvg -o command to get a list of currently varied-on volume

Free & Share & Open

314

Chapter 11 groups. Using this active VG list, we use a for loop to loop through each active VG and query for the STALE PVs: field using the lsvg $VG command to extract the number of stale PVs in each VG using both grep and awk. When any stale PVs are detected, the VG is added to the STALE_VG_LIST variable, all of the PVs in the VG are then added to the PV_LIST variable, specified by the lsvg -p $VG command. Next the STALE_ PV_COUNT variable is incremented by one for each PV using the math notation ((STALE_PV_COUNT = STALE_PV_COUNT + 1)). At this point we have a list of all of the volume groups that have stale physical volumes and a list of all of the PVs in all of the VGs that have stale PVs identified. If the STALE_PV_COUNT variable is zero, there are no stale disk partitions to report in the system for the currently varied-on volume groups. If the count is zero, we inform the user that no stale disk mirrors were found, and we exit the script with a return code of 0, zero. If no stale disk partitions exist, this shell script executes in seconds. If the count is greater than zero, then we inform the user that stale disk mirrors were found, and we continue to the next step, which is to query each PV in the $PV_LIST searching for stale disk partitions. To query each PV that is part of a VG that has stale PVs identified, we use a for loop to loop through each hdisk assigned to the $PV_LIST variable. Before we can query the disk, we need to ensure that the PV is in an active state. If the disk is inactive, then we cannot query that disk. In this section of the shell script we use the lspv $HDISK command within the for loop twice. The first time we are ensuring that the disk is active, and in the second step we query the disk for value of the STALE PARTITIONS: field. If the disk is found to be inactive, then we just add the disk to the INACTIVE_PV_LIST variable. If the disk is in an active state and the query detects any stale partitions, we add the hdisk to the STALE_PV_LIST variable. Notice in this section the if statement syntax that is used to check for inactive PVs before the disk query is initiated: if ! $(echo $INACTIVE_PV_LIST | grep $HDISK) >/dev/null 2>&1

The previous test ensures that the disk is not listed in the $INACTIVE_PV_LIST variable. The nice thing about using this syntax is that we use the if statement to check the return code of the enclosed command. We also negate the response so that we are testing for the disk not being listed in the variable by using the ! operator. To stop any screen output, the command is redirected to the bit bucket, and standard error is redirected to standard output, specified by the 2>&1 notation. Through the process of this for loop we populate the STALE_PV_LIST variable, which is a list of each of the active disks on the system that have stale disk partitions. We also keep a running count of the stale PPs found. In the next section we use the populated $STALE_PV_LIST variable to get a list of all of the logical volumes that are part of each disk in the stale disk list. In this step we use another for loop to loop through each stale PV and populate the STALE_LV_LIST variable using the lspv -l $PV command. Then we use this newly populated $STALE_LV_LIST to query each LV to find which ones have stale PPs. For this section we query each LV using the lslv $LV command and extract the value of the STALE

Monitor for Stale Disk Partitions
PP: field using a combination of grep and awk commands in a pipe. Each LV found to have at least one stale PP is added to the RESYNC_LV_LIST variable, which is used later to resync each of the LVs, if enabled, and in the log report. Now we use the list of inactive PVs, using the $INACTIVE_PV_LIST variable, to produce notification to the user of each inactive PV found on the system. We start with an if statement and test for the $INACTIVE_PV_LIST variable being NULL, or empty. If the variable is not NULL, then we loop through each PV in the list and issue a warning message to the user for each inactive PV. This information is also logged in the $LOGFILE using a pipe to the tee -a command to append to the $LOGFILE and display the information to the screen at the same time. In the next step, we give the user a list of each VG, PV, and LV that is affected by the stale disk partitions. After this notification is both logged and displayed we attempt to resync the mirrors at the LV level. Sometimes there are just one or two LVs on a PV that have stale disk partitions, so the LV is where we want to attempt to resync. We will attempt a resync only if the $ATTEMPT_RESYNC variable is initialized to TRUE. Any other value will cause this step to be skipped, but the user is notified that the resync option is disabled. If a resync is enabled, the syncvg -l $RESYNC_LV_LIST command is executed. The return code is checked for a zero value, indicating a successful resync operation. If the return code is not zero, you need to call IBM support and replace the disk before it goes dead on you. The steps involved in replacing a disk are beyond the scope of this book. We can also use the varyonvg command to resync only the stale partitions.

315

Other Options to Consider
As usual, any shell script can be improved, and this set of shell scripts is no exception.

SSA Disks
The ssaxlate command is used with a type of disk developed by IBM known as Serial Storage Architecture (SSA). The SSA disks not only use the hdisk# but also have an associated pdisk#. Normally the hdisk# and the pdisk# differ on the system. The ssaxlate command gives a cross-reference between the two disk representations. It is always a good idea to have this extra information if we are dealing with SSA disks, especially if you are replacing an SSA disk. To use the ssaxlate command, you need to know the specific hdisk# to translate to the corresponding pdisk#, or vice versa. As an example, we want to know what pdisk# translates to hdisk36. The command syntax to do the translation is shown here:
# ssaxlate -l hdisk36 pdisk32

In this example, hdisk36 translates to pdisk32. From this you can imply that hdisk0 through hdisk3 are not SSA disks. Usually the first few disks on an AIX system are

Free & Share & Open

316

Chapter 11
SCSI disk drives. You can also translate a pdisk# into the corresponding hdisk# by running the ssaxlate command against the pdisk#.

Log Files
In the first two shell scripts in this chapter we did not use a log file as we did in Method 3. It is always nice to have a log file to look at after the fact when you are running any type of system query. Creating a log file is a simple process of defining a variable to point to a filename that you want to use for a log file and appending output to the log file variable. If your system tends to fill up the /tmp filesystem, then I recommend creating a log directory, maybe in /usr/local/logs, or creating a separate filesystem just for log files. You can still have the mount point /usr/local/logs, or anything you want. If /tmp fills up, then you will not be able to write anything to the log file. You may also want to keep a month’s worth of log files to review in case of system problems. To do this you can add a date stamp as a filename extension and remove all files older than 30 days with the find command.

Automated Execution
You can make a cron entry in the root cron table to execute this shell script to automate running the script daily. A sample cron table entry is shown here:
05 23 * * * /usr/local/bin/stale_PP_mon.ksh >/dev/null 2>&1

The previous cron table entry will execute the stale_PP_mon.ksh shell script every day at 11:05 P.M., 365 days a year. The output is redirected to the bit bucket, but the log file will be created for review the next day.

Event Notification
If you use the previous cron table entry to execute the shell script every day, you may want to get some kind of notification by way of an email or a page. The easiest way is to email the log file to yourself every day. You can also modify the shell script to produce a very short message as a page. As an example, you could send one of the following text messages to an alphanumeric pager:
$THIS_HOST: stale PP check OK $THIS_HOST: stale PP check FAILED

These are short messages to get the point across, and you will know which machine the page came from.

Monitor for Stale Disk Partitions

317

Summary
In this chapter we looked at a logical progression of creating a shell script by starting at the basics. I hope you have gained at least some knowledge of the AIX Logical Volume Manager (LVM) through this experience. As you can see in this chapter, the first attempt to solve a challenge may not always be the best, or fastest, method; but this is how we learn. If we take these small steps and work up a full-blown shell script with all of the bells and whistles, we have learned a great deal. I know a lot of you do not work on AIX systems but this is still a valuable exercise. In the next chapter we look at some techniques of automating the ping process to ensure that the machines can communicate, at least at the lowest level of a ping. This is just another step toward being proactive and looking like gold. See you in the next chapter!

Free & Share & Open

CHAPTER

12
Automated Hosts Pinging with Notification

In every shop there is a critical need to keep the servers serving. For system availability, the quicker you know that a system is unreachable, the quicker you can act to resolve the problem and reduce company losses. At the lowest level of system access we can ping each machine in the “critical machine” list. If the ping works it will tell you if the network adapter is working, but it does not guarantee that the machine and applications are working. For this level of checks you need to actually access the application or operating system. In this chapter we are going to create a shell script that will ping hosts using a list of machines, which is stored in a separate file that is easily edited. Other options to this scenario include pinging all of the machines in the /etc/hosts file, using ftp to transfer a file, and querying the database, to name a few. Our interest in this chapter is to work at the lowest level and use the ping command to ensure that the machines are reachable from the network. When a machine is found unreachable we send notification to alert staff that the machine is down. Due to the fact that in some shops the network can become saturated with network traffic, we are going to add an extra level of testing on a failed ping test, which we will get into later in this chapter. But before we go any further let’s look at the command syntax for each of our operating systems (AIX, HP-UX, Linux, and Solaris) to see if we can find a command syntax that will produce the same output for all of the operating systems that we are working with.

319

Free & Share & Open

320

Chapter 12

Syntax
As always, we need the correct command syntax before we can write a shell script. Our goal is to find the command syntax for each operating system that produces the same output. For this shell script we want to ping each host multiple times to ensure that the node is reachable; the default is three pings. The standard output we want to produce on each OS is shown here.
# ping -c3 dino PING dino: (10.10.10.4): 56 data bytes 64 bytes from 10.10.10.4: icmp_seq=0 ttl=255 time=2 ms 64 bytes from 10.10.10.4: icmp_seq=1 ttl=255 time=1 ms 64 bytes from 10.10.10.4: icmp_seq=2 ttl=255 time=1 ms ----dino PING Statistics---3 packets transmitted, 3 packets received, 0% packet loss round-trip min/avg/max = 1/1/2 ms

This is the command, and the output is from an AIX machine. Notice the PING Statistics at the bottom of the command output, where I have highlighted 3 packets received. This is the output line that we are interested in for every operating system. Now, how do we produce the same output for each OS? Instead of leaving you in the dark I am just going to list each one in Table 12.1, showing you how to ping the host dino. In Table 12.1, notice that AIX and Linux have the same command syntax. For HP-UX and Solaris notice the two numbers, 56 and 3. The 56 specifies the packet size to send on each ping, and the 3 is the number of times to try to reach the host. For a packet size 56 is a standard packet, and we are not going to change from this standard. It is important to know the differences in command structure for each operating system because we are creating one shell script and we will ping each node using a function, which selects the correct command to execute based on the Unix flavor. To find the OS we use the uname command. Using the output of the uname command in a case statement we are assured that the correct command is executed on any of the four operating systems. This is really all we have for the syntax, but we need to do some checks and create some variables, so we are going to build the shell script around these commands listed in Table 12.1.
Table 12.1 Ping Command for Each Operating System PING COMMAND # ping -c3 dino # ping dino 56 3 # ping -c3 dino # ping -s dino 56 3

OPERATING SYSTEM AIX HP-UX Linux Solaris

Automated Hosts Pinging with Notification

321

Creating the Shell Script
In scripting this solution we want to add a couple of options for convenience. The first option is to have a means of stopping the pinging without interrupting the scheduled script execution, which is usually executed through a cron table entry. The second option is to have a means of stopping the notification for nodes that are unreachable. For each of these we can use a flag variable that must have a value of TRUE to enable the option. There are many times where you want to disable these two options, but the main reason is during a maintenance window when many of the machines are unreachable at the same time. If you have only one or two machines that are down, then commenting out the node name(s) in the ping.list file, which contains a list of nodes to ping, is preferable. You can also comment out the cron table entry to disable the test alogether. Now we need to define the pinging technique that we want to use. I like to use a two-level approach in checking for a system’s reachability. In a two-level testing scenario, when a node is unreachable we go to sleep for a few seconds and try the test again. We do this to eliminate “false positives” due to a heavy network load. This is a major concern at some shops where I have worked, and finger pointing back and forth between the network team and the Systems Administrators always happens, and I try to stay out of this argument. This second-level test adds just a few seconds to the testing window for each unreachable node. This is a relatively simple shell script to create, so keep reading!

Define the Variables
The first thing that we want to do in almost any shell script is to define the variables and files that are used in the script. We have already discussed two variables, which enable pinging and notification. For pinging we use the PINGHOSTS variable and MAILOUT as the variable to permit or disable notification. Additionally, for ease of testing we are going to typeset both of these variables to force all text assignments to these variables to uppercase, as shown here. typeset -u PINGHOSTS typeset -u MAILOUT PINGHOSTS=TRUE MAILOUT=TRUE

We can also typeset the variables and assign the values in the same step, as shown here. typeset -u PINGHOSTS=true typeset -u MAILOUT=true

Notice that I assign a lowercase “true” to both variables, but when you print or test the variables you will see that the assignments have been changed to uppercase characters.

Free & Share & Open

322

Chapter 12
# echo $MAILOUT TRUE

There are a few more variables that we also need to define, including PING_COUNT and PACKET_SIZE that specify the number of times to ping the target host and the packet size for each packet, which we discussed earlier. integer PING_COUNT=3 integer PACKET_SIZE=56

Notice the integer notation used to define these variables as integers. This notation produces the exact same results that the typeset -i command produces. Next we need the Unix flavor that this shell script is running. This shell script recognizes AIX, HP-UX, Linux, and Solaris. For this step we use the uname command, as shown here.
UNAME=$(uname)

In this UNAME assignment we used command substitution to assign the result of the uname command to the variable UNAME. The next two steps in this definition section involve defining the PINGFILE and MAILFILE file assignments. The PINGFILE contains a list of nodes that we want to ping. The shell script is expecting one node, or hostname, per line. If you no longer want to ping a node in the list file, then you can comment the node out using a pound sign (#). For this shell script I specified that the ping list is located in /usr/local/ bin/ping.list. Similarly, the MAILFILE has a list of email addresses that are to be notified when a node is not reachable. This email list is located in /usr/local/ bin/mail.list. The variable assignments are shown here.
PINGFILE=”/usr/local/bin/ping.list” # List of nodes to ping MAILFILE=”/usr/local/bin/mail.list” # List of persons to notify

For these two files we are going to check for a nonzero length file, which implies the file exists and its size is greater than zero bytes. If the $PINGFILE does not exist, then we need to send an ERROR message to the user and exit the shell script because we do not have a list of nodes to ping. If the $MAILFILE does not exist we are just going to notify the user that there will not be any email notification sent for unreachable nodes. We also need a file to hold the data that is emailed out when a node is unreachable. The file is located in /tmp/pingfile.out and is assigned to the PING_OUTFILE variable.
PING_OUTFILE=”/tmp/pingfile.out” # File for e-mailed notification >$PING_OUTFILE # Initialize to an empty file

Notice how we created an empty file by redirecting nothing to the file, which is pointed to by the $PING_OUTFILE variable. You could also use cat /dev/null to accomplish the same task, as shown here.

Automated Hosts Pinging with Notification cat /dev/null > $PING_OUTFILE

323

Next we need three variables that are to hold numeric values—at least we hope they are numeric. The first variable is called INTERVAL, and it contains a value specifying the number of seconds to sleep before trying to ping an unreachable node for the second time. I like to use three seconds. integer INTERVAL=”3” # Number of seconds to sleep between retries

As we discussed before, in our standard ping output we are interested in the PING Statistics line of output. Specifically, we want to extract the numeric value for the ”3 packets received”, which should be greater than zero if the node is reachable. To hold the value for the number of pings received back we need two variables, one for the first try and one for the second attempt, in case the node is unreachable the first time. These two variables are PINGSTAT and PINGSTAT2 and are initialized to NULL, as shown here.
PINGSTAT= PINGSTAT2= # Number of pings received back from pinging a node # Number of pings received back on the second try

The last variable we need to assign is the hostname of the machine that is running this script. We need the hostname because we may have two nodes pinging each node in case one pinging node fails. For this variable we again use command substitution, as shown here.
THISHOST=`hostname` # The hostname of this machine

Notice that this time we used the back tics (`command`) instead of the dollardouble parentheses method ($(command)) for command substitution. Both command substitution options produce the same result, which is yogi on this machine.

Creating a Trap
To start out our shell script we are going to set a trap, which allows us to take some kind of action when an exit signal is captured, such as a user pressing CTRL-C. We can capture most exit signals except for kill -9. The only action that we want to take in this shell script is to inform the user that the shell script has detected an exit signal and the script is exiting. This trap is added in this shell script so that you get used to putting traps in all of your shell scripts. We are going to capture exit signals 1, 2, 3, 5, and 15 only. You can add many more, but it is overkill in this case. For a complete list of signals use the kill -l (-ell) command. The command syntax for the trap is shown here. trap ‘echo “\nExiting on a trapped signal...\n”;exit 1’ 1 2 3 5 15

Using this trap command statement, the following message is displayed before the shell script exits with exit signal 1.
Exiting on a trapped signal...

Free & Share & Open

324

Chapter 12

The Whole Shell Script
We have all of the initializations complete and know what the ping syntax is for each operating system, so let’s look at the whole shell script and cover some other issues at the end of Listing 12.1. Pay close attention to the boldface text.
#!/usr/bin/ksh # # # SCRIPT: pingnodes.ksh # # AUTHOR: Randy Michael # # DATE: 02-20-2001 # # PURPOSE: This script is used to ping a list of nodes and # send email notification (or alphanumeric page) of any unreachable # nodes. # # # REV: 1.0.A # # REV.LIST: # # # set -x # Uncomment to debug this script # set -n # Uncomment to check command syntax without any execution # ####################################################### # Set a trap and clean up before a trapped exit... # REMEMBER: you CANNOT trap “kill -9” trap ‘echo “\n\nExiting on trapped signal...\n” \ ;exit 1’ 1 2 3 15 ####################################################### # Define and initialize variables here... PING_COUNT=”3” PACKET_SIZE=”56” # The number of times to ping each node # Packet size of each ping # Always use the UPPERCASE value for $PINGNODES # To enable or disable pinging FROM this node -

typeset -u PINGNODES PINGNODES=”TRUE” “TRUE” typeset -u MAILOUT

# Always use the UPPERCASE value for $MAILOUT

Listing 12.1 pingnodes.ksh shell script listing.

Automated Hosts Pinging with Notification

325

MAILOUT=”TRUE” events UNAME=$(uname)

# TRUE enables outbound mail notification of

# Get the Unix flavor of this machine

PINGFILE=”/usr/local/bin/ping.list” # List of nodes to ping if [ -s $PINGFILE ] then # Ping all nodes in the list that are not commented out and blank PINGLIST=$(cat $PINGFILE | grep -v ‘^#’) else echo “\nERROR: Missing file - $PINGFILE” echo “\nList of nodes to ping is unknown...EXITING...\n” exit 2 fi MAILFILE=”/usr/local/bin/mail.list” # List of persons to notify if [ -s $MAILFILE ] then # Ping all nodes in the list that are not commented out and blank MAILLIST=$(cat $MAILFILE | egrep -v ‘^#’) else echo “\nERROR: Missing file - $MAILFILE” echo “\nList of persons to notify is unknown...\n” echo “No one will be notified of unreachable nodes...\n” fi PING_OUTFILE=”/tmp/pingfile.out” # File for emailed notification >$PING_OUTFILE # Initialize to an empty file integer INTERVAL=”3” # Number of seconds to sleep between retries # Initialize the next two variables to NULL PINGSTAT= PINGSTAT2= # Number of pings received back from pinging a node # Number of pings received back on the second try # The hostname of this machine

THISHOST=`hostname`

######################################################## ############ DEFINE FUNCTIONS HERE ##################### ######################################################## function ping_host {

Listing 12.1 pingnodes.ksh shell script listing. (continues)

Free & Share & Open

326

Chapter 12

# This function pings a single node based on the Unix flavor # set -x # Uncomment to debug this function # set -n # Uncomment to check the syntax without any execution # Look for exactly one argument, the host to ping if (( $# != 1 )) then echo “\nERROR: Incorrect number of arguments - $#” echo “ Expecting exactly one argument\n” echo “\t...EXITING...\n” exit 1 fi HOST=$1 # Grab the host to ping from ARG1. # This next case statement executes the correct ping # command based on the Unix flavor case $UNAME in AIX|Linux) ping -c${PING_COUNT} $HOST 2>/dev/null ;; HP-UX) ping $HOST $PACKET_SIZE $PING_COUNT 2>/dev/null ;; SunOS) ping -s $HOST $PACKET_SIZE $PING_COUNT 2>/dev/null ;; *) echo “\nERROR: Unsupported Operating System - $(uname)” echo “\n\t...EXITING...\n” exit 1 esac } ####################################################### function ping_nodes { ####################################################### # # Ping the other systems check # # This can be disabled if you do not want every node to be pinging all

Listing 12.1 pingnodes.ksh shell script listing. (continued)

Automated Hosts Pinging with Notification

327

# # # # # #

of the other nodes. It is not necessary for all nodes to ping all other nodes although you do want more than one node doing the pinging just in case the pinging node is down. To activate pinging the “$PINGNODES” variable must be set to “TRUE”. Any other value will disable pinging from this node.

# set -x # Uncomment to debug this function # set -n # Uncomment to check command syntax without any execution if [[ $PINGNODES = “TRUE” ]] then echo # Add a single line to the output # Loop through each node in the $PINGLIST for HOSTPINGING in $(echo $PINGLIST) # Spaces between nodes in the # list are assumed do # Inform the user what is going on echo “Pinging --> ${HOSTPINGING}...\c” # If the pings received back is equal to “0” then you have a # problem. # Ping $PING_COUNT times, extract the value for the pings # received back.

PINGSTAT=$(ping_host $HOSTPINGING | grep transmitted \ | awk ‘{print $4}’) # If the value of $PINGSTAT is NULL, then the node is # unknown to this host if [[ -z “$PINGSTAT” && “$PINGSTAT” = ‘’ ]] then echo “Unknown host” continue fi if (( PINGSTAT == 0 )) then # Let’s do it again to make sure it really is unreachable echo “Unreachable...Trying one more time...\c” sleep $INTERVAL

Listing 12.1 pingnodes.ksh shell script listing. (continues)

Free & Share & Open

328

Chapter 12

PINGSTAT2=$(ping_host $HOSTPINGING | grep transmitted \ | awk ‘{print $4}’) if (( PINGSTAT2 == 0 )) then # It REALLY IS unreachable...Notify!! echo “Unreachable” echo “Unable to ping $HOSTPINGING from $THISHOST” \ | tee -a $PING_OUTFILE else echo “OK” fi else echo “OK” fi done fi } ###################################################### function send_notification { if [ -s $PING_OUTFILE -a “$MAILOUT” = “TRUE” ]; then case $UNAME in AIX|HP-UX|Linux) SENDMAIL=”/usr/sbin/sendmail” ;; SunOS) SENDMAIL=”/usr/lib/sendmail” ;; esac echo “\nSending e-mail notification” $SENDMAIL -f randy@$THISHOST $MAILLIST < $PING_OUTFILE fi } ################################################## ############ START of MAIN ####################### ##################################################

ping_nodes send_notification

# End of script

Listing 12.1 pingnodes.ksh shell script listing. (continued)

Automated Hosts Pinging with Notification
Now we get to the fun stuff! Let’s start out with the three functions because they do all of the work anyway. The first function is ping_host. The idea here is to set up a case statement, and based on the response from the uname command, which was assigned to the UNAME variable in the definitions section, we execute the specific ping command for the particular Unix flavor. If an unlisted Unix flavor is given, an ERROR message is given to the user, and this shell script exits with a return code 1. We must do this because we have no idea what the correct syntax for a ping command should be for an unknown operating system. The ping_host function is called from the ping_nodes function on every loop iteration. Inside the ping_nodes function we first ensure that the $PINGNODES variable is set to TRUE; otherwise, the pinging of nodes is disabled. We use the $PINGFILE file to load a variable, PINGLIST, with a list of nodes that we want to ping. This extra step is done to give the user the ability to comment out specific node(s) in the $PINGFILE. Without this ability you would leave the user in a state of annoyance for all of the notifications because of a single node being down for a period of time. The command to strip out the commented lines and leave the remaining nodes in the list is shown here.
PINGLIST=$(cat $PINGFILE | grep -v ‘^#’)

329

Notice how this command substitution works. We cat the $PINGFILE and pipe the output to a grep command. In the grep part of the statement we use the -v switch. The -v switch tells grep to list everything except for the following pattern, which is “^#” in this case. Now let’s look at the ^# part. When you put a carat character (^) in front of a pattern in this grep statement, we are ignoring any line that begins with a pound sign (#). The carat (^) means begins with. A for loop is started using the $PINGLIST variable as a list, which contains each node in the /usr/local/bin/ping.list file that is not commented out. For each node in the listing we echo to the screen the target node name and call the ping_host function inside of a command substitution statement on each loop iteration, which is shown here. echo “Pinging --> ${HOSTPINGING}...\c” PINGSTAT=$(ping_host $HOSTPINGING | grep transmitted | awk ‘{print $4}’)

For each node in the $PINGLIST the echo statement and the command substitution statement are executed. There are three possible results for the command substitution statement, and we test for two; the last one is assumed. (1) The PINGSTAT value is 0, zero. If the packets received are 0, zero, then we sleep for $INTERVAL seconds and try to reach the node again, this time assigning the packets received to the PINGSTAT2 variable. (2) The PINGSTAT value is NULL. This results when you try to ping a node that is unknown to the system. In this case we echo to the screen Unknown host and continue to the next node in the list. (3) The PINGSTAT value is nonzero and nonNULL, which means that the ping was successful. Please study each of these tests in the ping_nodes function.

Free & Share & Open

330

Chapter 12
Notice the tests used in the if statements. Each of these is a mathematical test so we use the double parentheses method of testing, as shown here. if (( PINGSTAT == 0 ))

There are two things to notice in this if statement. The first is that there is no dollar sign ($) in front of the PINGSTAT variable. The dollar sign is not needed in a mathematical test when using the double parentheses method because the shell assumes that any nonnumeric string is a variable for this type of mathematical test. I have had cases where I added the dollar sign ($) in front of the variable, and it took me four days to figure out why the script was failing. In other cases I have seen the dollar sign used and the script worked without error. I always remove the dollar sign, just in case. This problem is extremely hard to find should an error occur. The second point I want to make in the previous if statement is the use of the double equal signs (==). Using this type of mathematical test, a single equal sign is an assignment, not an equality test. This sounds a little strange, but you can actually assign a value to a variable in a test. To test for equality, always use double equal signs (==) with this test method. The last function in this shell script is the send_notification function. This function is used to send an email notification to each address listed in the /usr/local/bin/mail.list file, which is pointed to by the MAILFILE variable. Before attempting any notification the function tests to see if the $PING_OUTFILE file has anything in it or if its size is greater than zero bytes. The second test is to ensure that the MAILOUT variable is set to TRUE. If the $PING_OUTFILE has some data and the MAILOUT variable is set to TRUE, then the function will attempt to notify each email address in the $MAILFILE. In the send_notification function notice that I am using the sendmail command, as opposed to the mail or mailx commands. I use the sendmail command because I worked at a shop where I had a lot of trouble getting mail through the firewall because I was sending the mail as root. I found a solution by using the sendmail command because I can specify a valid nonroot user as the person who sent the email. The command I use is shown here. sendmail -f randy@$THISHOST $MAILLIST < $PING_OUTFILE

In this statement the -f specifies who is sending the e-mail. The $MAILLIST is the list of persons who should receive the email, and the < $PING_OUTFILE input redirection is the body text of the email, which is stored in a file. I still have one little problem, though. The sendmail command is not always located in the same directory, and sometimes it is not in the $PATH. On AIX, HP-UX, and Linux the sendmail command is located in /usr/sbin. On Solaris the sendmail command is located in the /usr/lib directory. To get around this little problem we need a little case statement that utilizes the $UNAME variable that we used in the ping_host function. With a little modification we have the function shown in Listing 12.2.

Automated Hosts Pinging with Notification

331

function send_notification { if [ -s $PING_OUTFILE -a “$MAILOUT” = “TRUE” ]; then case $UNAME in AIX|HP-UX|Linux) SENDMAIL=”/usr/sbin/sendmail” ;; SunOS) SENDMAIL=”/usr/lib/sendmail” ;; esac echo “\nSending e-mail notification” $SENDMAIL -f randy@$THISHOST $MAILLIST < $PING_OUTFILE fi }

Listing 12.2 send_notification function listing.

Notice that we used a single line for AIX, HP-UX, and Linux in the case statement. At the end of the function we use the $SENDMAIL variable to point to the correct full path of the sendmail command for the specific operating system. Let’s not forget to look at the pingnodes.ksh shell script in action! In the following output, shown in Listing 12.3, the node dino is unknown to the system, and the mrranger node is powered down so there is no response from the ping to the system.
# ./pinghostfile.ksh.new Pinging --> yogi...OK Pinging --> bambam...OK Pinging --> booboo...OK Pinging --> dino...Unknown host Pinging --> wilma...OK Pinging --> mrranger...Unreachable...Trying one more time...Unreachable Unable to ping mrranger from yogi Sending e-mail notification

Listing 12.3 pingnodes.ksh shell script in action.

From the output in Listing 12.3, notice the result of pinging the node dino. I commented out the hostname dino in the /etc/hosts file. By doing so I made the node

Free & Share & Open

332

Chapter 12 unknown to the system because DNS is not configured on this system. The mrranger node is powered down so it is known but not reachable. Notice the difference in the outputs for these two similar, but very different, situations. Please study the code related to both of these tests in the ping_nodes function.

Other Options to Consider
As always, we can improve on any shell script, and this one is no exception. I have listed some options that you may want to consider.

$PINGLIST Variable Length Limit Problem
In this scripting solution we gave the user the capability to comment out specific nodes in the $PINGFILE. We assigned the list of nodes, which is a list without the comments, to a variable. This is fine for a relatively short list of nodes, but a problem arises when the maximum variable length, which is usually 2048 characters, is exceeded. If you have a long list of nodes that you want to ping and you notice that the script never gets to the end of the ping list, you have a problem. Or if you see a funny-looking node name, which is probably a hostname that has been cut off by the variable limit and associated with a system error message, then you have a problem. To resolve this issue, define a new file to point to the PINGLIST variable, and then we will use the file to store the ping list data instead of a variable. To use PINGLIST as a file, add/ change the following lines: ADD THIS LINE:
PINGLIST=/tmp/pinglist.out

CHANGE THIS LINE:
PINGLIST=$(cat $PINGFILE | grep -v ‘^#’)

TO THIS LINE: cat $PINGFILE | grep -v ‘^#’ > $PINGLIST

CHANGE THIS LINE: for HOSTPINGING in $(echo $PINGLIST)

TO THIS LINE: for HOSTPINGING in $(cat $PINGLIST)

Automated Hosts Pinging with Notification
Using the file to store the ping list data changes the limit to the maximum file size that the system supports or when the filesystem fills up, which should be plenty of space for anyone. This modified shell script is located on this book’s companion Web site. The script name is pingnodes_using_a_file.ksh.

333

Ping the /etc/hosts File Instead of a List File
This may be overkill for any large shop, but it is easy to modify the shell script to accomplish this task. You want to make the following change to the shell script after completing the tasks in the previous section “$PINGLIST Variable Length Limit Problem” to the shell script shown in Listing 12.1. CHANGE THESE LINES: if [ -s $PINGFILE ] then PINGLIST=$(cat $PINGFILE | grep -v ‘^#’)

TO THESE LINES: if [ -s /etc/hosts ] then # Ping all nodes in the /etc/hosts file cat /etc/hosts | sed /^#/d | sed /^$/d | grep -v 127.0.0.1 \ | awk ‘{print $2}’ > $PINGLIST

In this changed code we cat the /etc/hosts file and pipe the output to a sed statement, sed /^#/d. This sed statement removes every line in the /etc/hosts file that begins with a pound sign (#). The output of this sed statement is then piped to another sed statement, sed /^$/d, which removes all of the blank lines in the /etc/hosts file (the blank lines are specified by the ^$). This sed output is sent to a grep command that removes the loopback address from the list. Finally, the remaining output is piped to an awk statement that extracts the hostname out of the second field. The resulting output is redirected to the $PINGLIST file. This modified shell script to ping the /etc/hosts file is included on the Web site that accompanies the book. The filename is pinghostsfile.ksh.

Logging
I have not added any logging capability to this shell script. Adding a log file, in addition to user notification, can help you find trends of when nodes are unreachable. Adding a log file is not too difficult to do. The first step is to define a unique log filename in the definitions section and assign the filename to a variable, maybe LOGFILE. In the script test for the existence of the file, using a test similar to the following statement will work.

Free & Share & Open

334

Chapter 12
ADD THESE LINES:
LOGPATH=/usr/local/log LOGFILE=${LOGPATH}/pingnodes.log if [ ! -s $LOGFILE ] then if [ ! -d $LOGPATH ] then echo “\nCreating directory ==> $LOGPATH\c” mkdir /usr/local/log if (( $? != 0 )) then echo “\nUnable to create the $LOGPATH directory...EXITING \n” exit 1 fi chown $USER /usr/local/log chmod 755 $LOGPATH echo fi echo “\nCreating Logfile ==> $LOGFILE\c” cp /dev/null > $LOGFILE chown $USER $LOGFILE echo fi

After adding these lines of code, use the tee -a $LOGFILE command in a pipe to both display the text on the screen and log the data in the $LOGFILE.

Notification of “Unknown Host”
You may want to add notification, and maybe logging too, for nodes that are not known to the system. This usually occurs when the machine cannot resolve the node name into an IP address. This can be caused by the node not being listed in the /etc/hosts file or failure of the DNS lookup. Check both conditions when you get the Unknown host message. Currently, this shell script only echoes this information to the screen. You may want to add this message to the notification.

Notification Method
In this shell script we use email notification. I like email notification, but if you have a network failure this is not going to help you. To get around the network down problem with email, you may want to set up a modem, for dial-out only, to dial your alphanumeric pager number and leave you a message. At least you will always get the message. I have had times, though, when I received the message two hours later due to a message overflow to the modem.

Automated Hosts Pinging with Notification
You may just want to change the notification to another method, such as SNMP traps. If you execute this shell script from an enterprise management tool, then the response required back to the program is usually an SNMP trap. Refer to the documentation of the program you are using for details.

335

Automated Execution Using a Cron Table Entry
I know you do not want to execute this shell script from the command line every 15 minutes yourself! I use a root cron table entry to execute this shell script every 15 minutes, 24 hours a day, Monday through Saturday, and 8:00 A.M. to midnight on Sunday; of course, this requires two cron table entries. Because weekly backups and reboots happen early Sunday morning, I do not want to be awakened every Sunday morning when a machine reboots, so I have a special cron entry for Sunday. Both root cron table entries shown execute this script every 15 minutes.
5,20,35,50 * * * 1-6 /usr/local/bin/pingnodes.ksh >/dev/null 2>&1 5,20,35,50 8-23 * * 0 /usr/local/bin/pingnodes.ksh &1

The first entry executes the pingnodes.ksh shell script at 5, 20, 35, and 50 minutes of every hour from Monday through Saturday. The second entry executes the ping-nodes.ksh shell script at 5, 20, 35, and 50 minutes from 8:00 A.M. until 11:59 P.M., with the last ping test running at 11:50 P.M. Sunday night.

Summary
In this chapter we took a different approach than that of some other shell scripts in this book. Instead of creating a different function for each operating system, we created a single shell script and then used a separate function to execute the correct command syntax for the specific operating system. The uname command is a very useful tool for shell scripting solutions for various Unix flavors in a single shell script. I hope you enjoyed this chapter. I think we covered some unique ways to solve the scripting problems that arise when programming for multiple Unix flavors in the same script. In the next chapter we will dive into the task of taking a system snapshot. The idea is to get a point-in-time system configuration for later comparison if a system problem has you puzzled. See you in the next chapter!

Free & Share & Open

CHAPTER

13
Taking a System Snapshot

Have you ever rebooted a system and it came up in an unusual state? Any time you reboot a system you run a risk that the system will not come back up properly. When problems arise it is nice to have before and after pictures of the state of the machine. In this chapter we are going to look at some options for shell scripts that execute a series of commands to take a snapshot of the state of the machine. Some of the things to consider for this system snapshot include filesystems that are mounted, NFS mounts, processes that are running, network statistics and configuration, and a list of defined system resources, just to name a few. This is different from gathering a snapshot of performance statistics, which is gathered over a period of time. All we are looking for is system configuration data and the system’s state at a point in time, specifically before the system is rebooted or when it is running in a normal state with all of the applications running properly. With this information captured before a system reboot, you have a better chance of fixing a reboot problem quickly and reducing down time. I like to store snapshot information in a directory called /usr/local/reboot with the command names used for filenames. For this shell script all of the system information is stored in a single file with a section header added for each command output. Overall, this is not a difficult shell script to write, but gathering the list of commands that you want to run can sometimes be a challenge. For example, if you want to gather an application’s configuration you need to find the commands that will produce the desired output. I always prefer having too much information, rather than not enough information, to troubleshoot a problem.

337

Free & Share & Open

338

Chapter 13
In this chapter I have put together a list of commands and created a bunch of functions to execute in the shell script. The commands selected are the most critical for troubleshooting an AIX machine; however, you will need to tailor this set of commands to suit your particular needs, operating system, and environment. Every shop is different, but they are all the same in some sense, especially when it comes to troubleshooting a problem. Let’s look at some commands and the syntax that is required.

Syntax
As always, we need the commands and the proper syntax for these commands before we can write a shell script. The commands presented in this section are just a sample of the information that you can gather from the system. This set of commands is for an AIX system, but most apply to other Unix flavors with modified syntax. The list of AIX commands is shown in Listing 13.1.
# Hostname of the machine hostname OR uname -n # Unix flavor uname -s # AIX OS version oslevel # AIX maintenance level patch set instfix -i | grep AIX_ML OR oslevel -r # Time zone for this system cat /etc/environment | grep TZ | awk -F’=’ ‘{print $2}’ # Real memory in the system echo “$(bootinfo -r)KB” OR lsattr -El -a realmem | awk ‘{print $2}’ # Machine type/architecture uname -M OR - Depending on the architecture uname -p # List of defined system devices lsdev -C # Long directory listing of /dev ls -l /dev # List of all defined disks lsdev -Cc disk # List of all defined pdisks for SSA disks lsdev -Cc pdisk # List of defined tape drives

Listing 13.1 System snapshot commands for AIX.

Taking a System Snapshot

339

lsdev -Cc tape # List of defined CD-ROMs lsdev -Cc cdrom # List of all defined adapters lsdev -Cc adapter # List of network routes netstat -rn # Network adapter statistics netstat -i # Filesystem Statistics df -k AND mount # List of defined Volume Groups lsvg | sort -r # List of varied-on Volume Groups lsvg -o | sort -r # List of Logical Volumes in each Volume Group for VG in $(lsvg -o | sort -r) do lsvg -l $VG done # Paging space definitions and usage lsps -a AND lsps -s # List of all hdisks in the system lspv # Disk drives listed by Volume Group assignment for VG in $(lsvg -o | sort -r) do lsvg -p $VG done # List the HACMP configuration, if installed if [ -x /usr/sbin/cluster/utilities/cllsif ] then /usr/sbin/cluster/utilities/cllsif echo “\n” fi if [ -x /usr/sbin/cluster/utilities/clshowres ] then /usr/sbin/cluster/utilities/clshowres fi # List of all defined printers lpstat -W | tail +3 AND cat /etc/qconfig

Listing 13.1 System snapshot commands for AIX. (continues)

Free & Share & Open

340

Chapter 13

# List of active processes ps -ef # Show SNA configuration, if installed sna -d s if (($? != 0)) then lssrc -s sna -l fi # List of udp and x25 processes, if any ps -ef | egrep ‘udp|x25’ | grep -v grep # Short listing of the system configuration lscfg # Long listing of the system configuration lscfg -vp # List of all system installed filesets lslpp -L # List of broken or inconsistant filesets lppchk -v 2>&1 # List of the last 100 users to log in to the system last | tail -100

Listing 13.1 System snapshot commands for AIX. (continued)

As you can see in Listing 13.1, we can add anything that you want to the snapshot shell script to get as much detail as needed to troubleshoot a problem. Every environment is different, so this list of commands should be modified, or added to, to suit the needs of your shop. Additional tests include a list of databases that are running, application configurations, specific application processes that are critical, and a ping list of machines that are critical to the operation of any applications. You can add anything that you want or need here. Always try to gather more information than you think you may need to troubleshoot a problem. Using this snapshot technique allows us to go back and look at what the system looked like under normal conditions and load. By looking at the snapshot script output file, the problem usually stands out when comparing it to the currently running system that has a problem.

Creating the Shell Script
For this shell script we are going to take the commands shown in Listing 13.1 and create a function for each one. Using functions greatly simplifies both creating and modifying the entire shell script. When we want to add a new test, or configuration output, we just create a new function and add the function-name in the main body of the shell script exactly where we want it to run. In this shell script all of the function definitions use the C-like function statement, as shown here.

Taking a System Snapshot get_last_logins () { Commands to execute }

341

A lot of script programmers like this function definition technique. I prefer defining a function using the function statement method, as shown here. function get_last_logins { Commands to execute }

This last method of defining a function is more intuitive to understand for the people who will follow in your footsteps and modify this shell script. I hope you noticed the use of the word will in the last sentence. No matter what the shell script does, there is always someone who will come along, after you have moved on to bigger and better things, who will modify the shell script. It is usually not because there is a problem with the script coding, but more likely a need for added functionality. For the people who follow me, I like to make sure that the shell script is easy to follow and understand. Use your own judgment and preference when defining functions in a shell script; just be consistent. Because we have all of the commands listed in Listing 13.1 let’s look at the entire shell script in Listing 13.2 and see how we created all of these functions.
#!/bin/ksh # # SCRIPT: AIXsysconfig.ksh # AUTHOR: Randy Michael # REV: 2.1.P # DATE: 06/14/2002 # # PLATFORM: AIX only # # PURPOSE: Take a snapshot of the system for later comparision in the # event of system problems. All data is stored in # /usr/local/reboot in the file defined to the $SYSINFO_FILE # variable below. # # # REV LIST: # 7/11/2002: Changed this script to use a single output file # that receives data from a series of commands # within a bunch of functions. # #

Listing 13.2 AIXsysconfig.ksh shell script listing. (continues)

Free & Share & Open

342

Chapter 13

# set -x # Uncomment to debug this script # set -n # Uncomment to verify command syntax without execution # ################################################# ######### DEFINE VARIABLES HERE ################# ################################################# THISHOST=$(/usr/bin/hostname) DATETIME=$(/usr/bin/date +%m%d%y_%H%M%S) WORKDIR=”/usr/local/reboot” SYSINFO_FILE=”${WORKDIR}/sys_snapshot.${THISHOST}.$DATETIME” ################################################# ############ DEFINE FUNCTIONS HERE ############## ################################################# get_host () { # Hostname of this machine hostname # uname -n works too } ################################################# get_OS () { # Operating System - AIX or exit uname -s } ################################################# get_OS_level () { # Query for the operating system release and version level oslevel } ################################################# get_ML_for_AIX () { # Query the system for the maintenance level patch set instfix -i | grep AIX_ML

Listing 13.2 AIXsysconfig.ksh shell script listing. (continued)

Taking a System Snapshot

343

echo “\n” oslevel -r } ################################################# get_TZ () { # Get the time zone that the system is operating in. cat /etc/environment | grep TZ | awk -F’=’ ‘{print $2}’ } ################################################# get_real_mem () { # Query the system for the total real memory echo “$(bootinfo -r)KB” # lsattr -El sys0 -a realmem | awk ‘{print $2}’ Works too } ################################################# get_arch () { # Query the system for the hardware architecture. Newer # machines use the -M switch, and the older Micro-Channel # architecture (MCA) machines use the -p option for # the “uname” command. ARCH=$(uname -M) if [[ -z “$ARCH” && “$ARCH” = ‘’ ]] then ARCH=$(uname -p) fi echo “$ARCH” } ################################################# get_devices () { # Query the system for all configured devices lsdev -C } #################################################

Listing 13.2 AIXsysconfig.ksh shell script listing. (continues)

Free & Share & Open

344

Chapter 13

get_long_devdir_listing () { # Long listing of the /dev directory. This shows the # device major and minor numbers and raw device ownership ls -l /dev } ################################################# get_defined_disks () { # List of all defined disks lsdev -Cc disk } ################################################# get_defined_pdisks () { # List of all defined pdisks for SSA disks lsdev -Cc pdisk } ################################################# get_tape_drives () { # Query the system for all configured tape drives lsdev -Cc tape } ################################################# get_cdrom () { # Query the system for all configured CD-ROM devices lsdev -Cc cdrom } ################################################# get_adapters () { # List all configured adapters in the system lsdev -Cc adapter }

Listing 13.2 AIXsysconfig.ksh shell script listing. (continued)

Taking a System Snapshot

345

################################################# get_routes () { # Save the network routes defined on the system netstat -rn } ################################################# get_netstats () { # Save the network adapter statistics netstat -i } ################################################# get_fs_stats () { # Save the file system statistics df -k echo “\n” mount } ################################################# get_VGs () { # List all defined Volume Groups lsvg | sort -r } ################################################# get_varied_on_VGs () { # List all varied-on Volume Groups lsvg -o | sort -r } ################################################# get_LV_info () {

Listing 13.2 AIXsysconfig.ksh shell script listing. (continues)

Free & Share & Open

346

Chapter 13

# List the Logical Volumes in each varied-on Volume Group for VG in $(get_varied_on_VGs) do lsvg -l $VG done } ################################################# get_paging_space () { # List the paging space definitions and usage lsps -a echo “\n” lsps -s } ################################################# get_disk_info () { # List of all “hdisk”s (hard drives) on the system lspv } ################################################# get_VG_disk_info () { # List disks by Volume Group assignment for VG in $(get_varied_on_VGs) do lsvg -p $VG done } ################################################# get_HACMP_info () { # If the System is running HACMP then save the # HACMP configuration if [ -x /usr/sbin/cluster/utilities/cllsif ] then /usr/sbin/cluster/utilities/cllsif echo “\n\n”

Listing 13.2 AIXsysconfig.ksh shell script listing. (continued)

Taking a System Snapshot

347

fi if [ -x /usr/sbin/cluster/utilities/clshowres ] then /usr/sbin/cluster/utilities/clshowres fi } ################################################# get_printer_info () { # Wide listing of all defined printers lpstat -W | tail +3 echo “\nPrint Queue Configuration File Listing\n” cat /etc/qconfig | grep -v ^* } ################################################# get_process_info () { # List of all active processes ps -ef } ################################################# get_sna_info () { # If the system is using SNA save the SNA configuration sna -d s # Syntax for 2.x SNA if (( $? != 0 )) then lssrc -s sna -l # must be SNA 1.x fi } ################################################# get_udp_x25_procs () { # Listing of all “udp” and “x25” processes, if # any are running ps -ef | egrep ‘udp|x25’ | grep -v grep } #################################################

Listing 13.2 AIXsysconfig.ksh shell script listing. (continues)

Free & Share & Open

348

Chapter 13

get_sys_cfg () { # Short listing of the system configuration lscfg } ################################################# get_long_sys_config () { # Long detailed listing of the system configuration lscfg -vp } ################################################# get_installed_filesets () { # Listing of all installed LPP filesets (system installed) lslpp -L } ################################################# check_for_broken_filesets () { # Check the system for broken filesets lppchk -v 2>&1 } ################################################# last_logins () { # List the last 100 system logins last | head -100 } ################################################# ############## START OF MAIN ################### ################################################# # Check for AIX as the operating system if [[ $(get_OS) != ‘AIX’ ]] then

Listing 13.2 AIXsysconfig.ksh shell script listing. (continued)

Taking a System Snapshot

349

echo “\nERROR: Incorrect operating system. This shell script is written for AIX.\n” echo “\n\t...EXITING...\n” exit 1 fi ################################################# # # Define the working directory and create this # directory if it does not exist. if [ ! -d $WORKDIR ] then mkdir -p $WORKDIR >/dev/null 2>&1 if (($? != 0)) then echo “\nERROR: Permissions do not allow you to create the $WORKDIR directory. This script must exit. Please create the $WORKDIR directory and execute this script again.\n” echo “\n\t...EXITING...\n” exit 2 fi fi ################################################# { # Everything enclosed between this opening bracket and the # later closing bracket is both displayed on the screen and # also saved in the log file defined as $SYSINFO_FILE.

echo “\n\n[ $(basename $0) - $(date) ]\n” echo “Saving system information for $THISHOST...” echo echo echo echo echo echo echo “\nSystem:\t\t\t$(get_host)” “Time Zone:\t\t$(get_TZ)” “Real Memory:\t\t$(get_real_mem)” “Machine Type:\t\t$(get_arch)” “Operating System:\t$(get_OS)” “OS Version Level:\t$(get_OS_level)” “\nCurrent OS Maintenance Level:\n$(get_ML_for_AIX)”

echo “\n#################################################\n” echo “Installed and Configured Devices\n” get_devices

Listing 13.2 AIXsysconfig.ksh shell script listing. (continues)

Free & Share & Open

350

Chapter 13

echo “\n#################################################\n” echo “Long Device Directory Listing - /dev\n” get_long_devdir_listing echo “\n#################################################\n” echo “\nSystem Defined Disks\n” get_defined_disks echo “\n#################################################\n” echo “\nSystem Defined SSA pdisks\n” get_defined_pdisks echo “\n#################################################\n” echo “System Tape Drives\n” get_tape_drives echo “\n#################################################\n” echo “System CD-ROM Drives\n” get_cdrom echo “\n#################################################\n” echo “Defined Adapters in the System\n” get_adapters echo “\n#################################################\n” echo “Network Routes\n” get_routes echo “\n#################################################\n” echo “Network Interface Statistics\n” get_netstats echo “\n#################################################\n” echo “Filesystem Statistics\n” get_fs_stats echo “\n#################################################\n” echo “Defined Volume Groups\n” get_VGs echo “\n#################################################\n” echo “Varied-on Volume Groups\n” get_varied_on_VGs echo “\n#################################################\n” echo “Logical Volume Information by Volume Group\n” get_LV_info echo “\n#################################################\n” echo “Paging Space Information\n” get_paging_space echo “\n#################################################\n” echo “Hard Disks Defined\n” get_disk_info echo “\n#################################################\n” echo “Volume Group Hard Drives\n” get_VG_disk_info echo “\n#################################################\n” echo “HACMP Configuration\n” get_HACMP_info

Listing 13.2 AIXsysconfig.ksh shell script listing. (continued)

Taking a System Snapshot

351

echo “\n#################################################\n” echo “Printer Information\n” get_printer_info echo “\n#################################################\n” echo “Active Process List\n” get_process_info echo “\n#################################################\n” echo “SNA Information\n” get_sna_info echo “\n#################################################\n” echo “x25 and udp Processes\n” get_udp_x25_procs echo “\n#################################################\n” echo “System Configuration Overview\n” get_sys_cfg echo “\n#################################################\n” echo “Detailed System Configuration\n” get_long_sys_config echo “\n#################################################\n” echo “System Installed Filesets\n” get_installed_filesets echo “\n#################################################\n” echo “Looking for Broken Filesets\n” check_for_broken_filesets echo “\n#################################################\n” echo “List of the last 100 users to log in to $THISHOST\n” last_logins echo “\n\nThis report is save in: $SYSINFO_FILE \n” # Send all output to both the screen and the $SYSINFO_FILE # using a pipe to the “tee -a” command” } | tee -a $SYSINFO_FILE

Listing 13.2 AIXsysconfig.ksh shell script listing. (continued)

As you can see in Listing 13.2, we have a lot of functions in this shell script. When I created these functions I tried to place each one in the order that I want to execute in the shell script. This is not necessary as long as you do not try to use a function before it is defined. Because a Korn shell script is interpreted, as opposed to compiled, the flow goes from the top to the bottom. It makes sense that you have to define a function in the code above where the function is used. If we slip up and the function is defined below where it is used, then we may or may not get an error message. Getting an error message depends on what the function is supposed to do and how the function is executed in the shell script.

Free & Share & Open

352

Chapter 13
From the top of the shell script in Listing 13.2 we first define the variables that we need. The hostname of the machine is always nice to know, and it is required for the report-file definition and in the report itself. Next we create a date/time stamp. This $DATATIME variable is used in the report-file definition as well. We want the date and time because this script may be executed more than once in a single day. Next we define the working directory. I like to use /usr/local/reboot, but you can use any directory that you want. Finally, we define the report-file, which is assigned to the $SYSINFO_FILE variable. The next section is where all of the functions are defined. Notice that some of these functions contain only a single command, and some have a bit more code. In a shell script like this one it is a good idea to place every command in a separate function. Using this method allows you to change the commands to a different operating system simply by editing some functions and leaving the basic shell script operation intact. There are too many functions in this shell script to go over them one at a time, but an output of this shell script is shown in Listing 13.3. For details on the specific AIX commands please refer to the AIX documentation and man pages on an AIX system. At START OF MAIN we begin the real work. The first step is to ensure that the operating system is AIX. If this shell script is executed on another Unix flavor, then a lot of the commands will fail. If a non-AIX Unix flavor is detected, then the user receives an error message and the script exits with a return code of 1, one. Step two is to test for the existence of the $WORKDIR directory, which is defined as /usr/local/reboot in this shell script. If the directory does not exist, an attempt is made to create the directory. Not all users will have permission to create a directory here. If the directory creation fails, then the user receives an error message and is asked to create the directory manually and run the shell script again. If the operating system is AIX and the $WORKDIR exists, then we create the reportfile and begin creating the report. Notice that the entire list of functions and commands for the report is enclosed in braces, { code }. Then, after the final brace, at the end of the shell script, all of the output is piped to the tee -a command. Using this pipe to the tee -a command allows the user to see the report as it is being created and the output is written to the $SYSINFO_FILE file. Enclosing all of the code for the report within the braces saves a lot of effort to get the output to the screen and to the report file. The basic syntax is shown here.
{ report command report command . . . report command } | tee -a $SYSINFO_FILE

Within the braces we start by setting up the report header information, which includes the hostname, time zone, real memory, machine type, operating system, operating system version, and the maintenance level patch set of the operating system version.

Taking a System Snapshot
When the header is complete then the script executes the functions listed in the DEFINE FUNCTIONS HERE section. As I stated before, I tried to define the functions in the order of execution. Before each function is executed, a line of hash marks is written out to separate each report section, and then some section header information is written for the specific task. At the end, and just before the ending brace, the report filename is shown to the user to indicate where the report file is located. Let’s take a look at an abbreviated report output in Listing 13.3.

353

[ AIXsysconfig.ksh - Thu Jul 25 09:46:58 EDT 2002 ] Saving system information for yogi... System: Time Zone: Real Memory: Machine Type: Operating System: OS Version Level: yogi EST5EDT 131072KB powerpc AIX 5.1.0.0

Current OS Maintenance Level: Not all filesets for 5.0.0.0_AIX_ML were found. Not all filesets for 5.1.0.0_AIX_ML were found. ################################################# Installed and Configured Devices sys0 sysplanar0 ioplanar0 sio0 hdisk0 hdisk1 rmt0 cd0 proc0 mem0 mem1 mem2 mem3 fd0 lvdd tty0 rootvg hd5 hd6 . Available Available Available Available Available Available Available Available Available Available Available Available Available Available Available Available Defined Defined Defined 00-00 00-00 00-00 00-00 00-00-0S-0,0 00-00-0S-1,0 00-00-0S-5,0 00-00-0S-6,0 00-00 00-0A 00-0B 00-0C 00-0D 00-00-0D-00 00-00-S1-00 System Object System Planar I/O Planar Standard I/O Planar 2.0 GB SCSI Disk Drive 2.0 GB SCSI Disk Drive 5.0 GB 8mm Tape Drive SCSI Multimedia CD-ROM Drive Processor 32 MB Memory SIMM 32 MB Memory SIMM 32 MB Memory SIMM 32 MB Memory SIMM Diskette Drive LVM Device Driver Asynchronous Terminal Volume group Logical volume Logical volume

Listing 13.3 AIXsysconfig.ksh shell script in action. (continues)

Free & Share & Open

354

Chapter 13

. . ################################################# Long Device Directory Listing - /dev total 24 crw-rw-rwcrw-rw-rwcrw-rw-rwcrw-rw-rwcrw-rw-rwcrw-rw-rwcrw-rw-rwcrw-rw-rwcrw-rw---crw-rw---crw-rw-rwbrw-rw---crw-rw-rwcrw-rw-rwcrw-rw-rwcrw------crw------crw------crw------crw-rw-rwcrw-rw-rwcrw-rw-rwcrw--w--wcrw-rw-rwcrw-rw-rwcrw-rw-rwcrw-rw-rw. . .

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

root root root root root root root root root root root root root root root root root root root root root root root root root root root

system system system system system system system system system system system system system system system system system system system system system system system system system system system

19, 19, 19, 19, 19, 19, 19, 19, 10, 10, 13, 10, 11, 13, 13, 7, 7, 7, 7, 5, 5, 1, 18, 18, 26, 26, 2,

0 1 2 3 4 5 6 7 0 10 14 10 0 15 30 0 1 3 2 0 1 0 0 1 0 1 3

Jun Mar Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jun Jul Jul Jul

23 29 26 26 26 26 26 26 26 29 26 29 26 26 26 26 26 26 26 26 26 24 24 23 26 26 26

15:23 13:49 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 2001 17:53 17:58 15:18 2001 2001 2001

rmt0 rmt0.1 rmt0.2 rmt0.3 rmt0.4 rmt0.5 rmt0.6 rmt0.7 rootvg rscripts_lv sad scripts_lv scsi0 slog spx sysdump sysdumpctl sysdumpfile sysdumpnull systrace systrctl tty tty0 tty1 ttyp0 ttyp1 zero

################################################# System Defined Disks hdisk0 Available 00-00-0S-0,0 2.0 GB SCSI Disk Drive hdisk1 Available 00-00-0S-1,0 2.0 GB SCSI Disk Drive #################################################

Listing 13.3 AIXsysconfig.ksh shell script in action. (continued)

Taking a System Snapshot

355

System Defined SSA pdisks

################################################# System Tape Drives rmt0 Available 00-00-0S-5,0 5.0 GB 8mm Tape Drive ################################################# System CD-ROM Drives cd0 Available 00-00-0S-6,0 SCSI Multimedia CD-ROM Drive ################################################# Defined Adapters in the System sio0 Available 00-00 Standard I/O Planar fda0 Available 00-00-0D Standard I/O Diskette Adapter sioka0 Available 00-00-0K Keyboard Adapter sa0 Available 00-00-S1 Standard I/O Serial Port 1 sa1 Available 00-00-S2 Standard I/O Serial Port 2 scsi0 Available 00-00-0S Standard SCSI I/O Controller siota0 Available 00-00-0T Tablet Adapter sioma0 Available 00-00-0M Mouse Adapter ppa0 Available 00-00-0P Standard I/O Parallel Port Adapter ent0 Available 00-03 Ethernet High-Performance LAN Adapter (8ef5) . . . ################################################# Network Routes Routing tables Destination Groups

Gateway

Flags

Refs

Use

If

PMTU Exp

Route Tree for Protocol Family 2 (Internet): default 10.10.10.2 UGc 0 10.10/16 10.10.10.1 U 37 127/8 127.0.0.1 U 5 Route Tree for Protocol Family 24 (Internet v6):

0 135807 264

en0 en0 lo0

-

-

Listing 13.3 AIXsysconfig.ksh shell script in action. (continues)

Free & Share & Open

356

Chapter 13

::1

::1

UH

0

0

lo0 16896

-

################################################# Network Interface Statistics Name en0 en0 lo0 lo0 lo0 Mtu 1500 1500 16896 16896 16896 Network link#2 10.10 link#1 127 ::1 Address Ipkts Ierrs 2.60.8c.2d.75.b1 112330 yogi 112330 28302 loopback 28302 28302 Opkts Oerrs 108697 108697 28304 28304 28304 Coll 0 0 0 0 0

0 0 0 0 0

0 0 0 0 0

################################################# Filesystem Statistics Filesystem 1024-blocks /dev/hd4 32768 /dev/hd2 1449984 /dev/hd9var 53248 /dev/hd3 106496 /dev/hd1 4096 /proc /dev/hd10opt 655360 /dev/scripts_lv 102400 /dev/lv_temp 409600 Free %Used 10924 67% 61680 96% 10568 81% 70184 35% 3892 5% 16460 98% 25296 76% 350456 15% Iused %Iused Mounted on 1854 12% / 40941 12% /usr 673 6% /var 223 1% /tmp 55 6% /home - /proc 16260 10% /opt 887 4% /scripts 26 1% /tmpfs

node mounted mounted over options -------- ------------ ------------ -----/dev/hd4 / jfs /dev/hd2 /usr jfs /dev/hd9var /var jfs /dev/hd3 /tmp jfs /dev/hd1 /home jfs /proc /proc procfs /dev/hd10opt /opt jfs /dev/scripts_lv /scripts jfs /dev/lv_temp /tmpfs jfs

vfs

date --------------rw,log=/dev/hd8 rw,log=/dev/hd8 rw,log=/dev/hd8 rw,log=/dev/hd8 rw,log=/dev/hd8 rw rw,log=/dev/hd8 rw,log=/dev/hd8 rw,log=/dev/hd8

-----------Jul 23 18:56 Jul 23 18:56 Jul 23 18:56 Jul 23 18:56 Jul 23 18:57 Jul 23 18:57 Jul 23 18:57 Jul 23 18:57 Jul 23 18:57

################################################# Defined Volume Groups rootvg

Listing 13.3 AIXsysconfig.ksh shell script in action. (continued)

Taking a System Snapshot

357

################################################# Varied-on Volume Groups rootvg ################################################# Logical Volume Information by Volume Group rootvg: LV NAME POINT hd5 hd6 hd8 hd4 hd2 hd9var hd3 hd1 hd10opt scripts_lv lv_temp

TYPE boot paging jfslog jfs jfs jfs jfs jfs jfs jfs jfs

LPs 2 84 1 8 354 13 26 1 160 25 100

PPs 2 84 1 8 354 13 26 1 160 25 100

PVs 1 1 1 1 2 2 1 1 2 1 1

LV STATE closed/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd open/syncd

MOUNT N/A N/A N/A / /usr /var /tmp /home /opt /scripts /tmpfs

################################################# Paging Space Information Page Space hd6 Physical Volume hdisk0 Volume Group Size rootvg 336MB %Used Active Auto 10 yes yes Type lv

Total Paging Space 336MB

Percent Used 10%

################################################# Hard Disks Defined hdisk0 hdisk1 00003677cf068b62 000125608a48c132 rootvg rootvg

################################################# Volume Group Hard Drives

Listing 13.3 AIXsysconfig.ksh shell script in action. (continues)

Free & Share & Open

358

Chapter 13

rootvg: PV_NAME hdisk0 hdisk1

PV STATE active active

TOTAL PPs 479 479

FREE PPs 0 184

FREE DISTRIBUTION 00..00..00..00..00 92..00..00..00..92

################################################# HACMP Configuration

################################################# Printer Information hp4 hp4-ps hp4-gl yogi_hp4_1 yogi_hp4_1ps lp0 lp0 lp0 lp0 lp0 READY READY READY READY READY

Print Queue Configuration File Listing hp4: device = lp0 lp0: file = /dev/lp0 header = never trailer = never access = both backend = /usr/lib/lpd/piobe hp4-ps: device = lp0 lp0: file = /dev/lp0 header = never trailer = never access = both backend = /usr/lib/lpd/piobe hp4-gl: device = lp0 lp0: file = /dev/lp0 header = never trailer = never access = both

Listing 13.3 AIXsysconfig.ksh shell script in action. (continued)

Taking a System Snapshot

359

backend = /usr/lib/lpd/piobe yogi_hp4_1: device = lp0 lp0: file = /dev/lp0 header = never trailer = never access = both backend = /usr/lib/lpd/piobe yogi_hp4_1ps: device = lp0 lp0: file = /dev/lp0 header = never trailer = never access = both backend = /usr/lib/lpd/piobe ################################################# Active Process List UID PID PPID C STIME root 1 0 0 Jul 23 root 1950 1 0 Jul 23 root 2672 1 0 Jul 23 root 3140 1 0 Jul 23 root 3642 4644 0 17:11:20 root 3882 1950 0 Jul 23 connections root 4168 1950 0 Jul 23 root 4388 1950 0 Jul 23 root 4644 1950 0 Jul 23 nobody 4906 5418 0 Jul 23 daemon 8798 1950 0 Jul 23 root 9034 1950 0 Jul 23 root 9296 1950 0 Jul 23 root 9554 1950 0 Jul 23 root 9814 1950 0 Jul 23 root 10336 1 0 Jul 23 root 10588 1950 0 Jul 23 root 10842 1 0 Jul 23 root 11360 1950 0 Jul 23 root 11616 1 0 Jul 23 /usr/lpp/diagnostics/bin/diagd root 16820 15772 0 17:11:39 TTY pts/0 TIME 0:17 0:00 0:00 2:04 0:00 0:04 0:00 0:00 0:00 0:01 0:00 0:00 0:00 0:00 0:00 0:00 0:00 0:02 0:00 0:00 CMD /etc/init /usr/sbin/srcmstr /usr/lib/errdemon /usr/sbin/syncd 60 rpc.ttdbserver 100083 1 sendmail: accepting /usr/sbin/syslogd /usr/sbin/portmap /usr/sbin/inetd /usr/sbin/tftpd -n /usr/sbin/rpc.statd /usr/sbin/biod 6 /usr/sbin/nfsd 3891 /usr/sbin/rpc.mountd /usr/sbin/rpc.lockd /usr/sbin/uprintfd qdaemon /usr/sbin/cron /usr/sbin/writesrv

0:03 dtfile

Listing 13.3 AIXsysconfig.ksh shell script in action. (continues)

Free & Share & Open

360

Chapter 13

root 17540 16538 . . .

0 21:16:59

pts/3

0:00 /usr/bin/ksh

################################################# SNA Information 0513-085 The sna Subsystem is not on file. ################################################# x25 and udp Processes

################################################# System Configuration Overview INSTALLED RESOURCE LIST The following resources are installed on the machine. +/- = Added or deleted from Resource List. * = Diagnostic support not available. * + + + + + + + + + + + * . . . sys0 sysplanar0 ioplanar0 hdisk0 hdisk1 rmt0 cd0 proc0 mem0 mem1 mem2 mem3 sysunit0 00-00 00-00 00-00 00-00-0S-0,0 00-00-0S-1,0 00-00-0S-5,0 00-00-0S-6,0 00-00 00-0A 00-0B 00-0C 00-0D 00-00 System Object System Planar I/O Planar 2.0 GB SCSI Disk Drive 2.0 GB SCSI Disk Drive 5.0 GB 8mm Tape Drive SCSI Multimedia CD-ROM Drive (650 Processor 32 MB Memory SIMM 32 MB Memory SIMM 32 MB Memory SIMM 32 MB Memory SIMM System Unit

################################################# Detailed System Configuration

Listing 13.3 AIXsysconfig.ksh shell script in action. (continued)

Taking a System Snapshot

361

INSTALLED RESOURCE LIST WITH VPD The following resources are installed on your machine. sys0 sysplanar0 00-00 00-00 System Object System Planar

Part Number.................065G8317 EC Level....................00D28027 Processor Identification....00012560 ROS Level and ID............IPLVER1.3 LVL3.01,065G8318 Processor Component ID......0800004800000050 Device Specific.(Z0)........000000 Device Specific.(Z1)........000000 Device Specific.(Z2)........000000 Device Specific.(Z3)........000000 Device Specific.(Z4)........000000 Device Specific.(Z5)........000000 Device Specific.(Z6)........000000 Device Specific.(Z7)........000000 Device Specific.(Z8)........000000 Device Specific.(Z9)........000000 ROS Level and ID............OCS(00000C54) ROS Level and ID............SEEDS(28040203) hdisk0 00-00-0S-0,0 2.0 GB SCSI Disk Drive

Manufacturer................IBMRISC Machine Type and Model......0664M1H Part Number.................86F0101 ROS Level and ID............5 5A Serial Number...............00221833 EC Level....................895186 FRU Number..................86F0118 Device Specific.(Z0)........000002029F00001E Device Specific.(Z1)........75G3644 Device Specific.(Z2)........0983 Device Specific.(Z3)........95123 Device Specific.(Z4)........0002 Device Specific.(Z5)........22 Device Specific.(Z6)........895172 rmt0 00-00-0S-5,0 5.0 GB 8mm Tape Drive

Manufacturer................EXABYTE

Listing 13.3 AIXsysconfig.ksh shell script in action. (continues)

Free & Share & Open

362

Chapter 13

Machine Type and Model......IBM-8505 Device Specific.(Z1)........807A Part Number.................8191044 Serial Number...............082737 Device Specific.(LI)........00000001 EC Level....................D48098 FRU Number..................59H3159 Device Specific.(Z0)........0180020283000010 cd0 00-00-0S-6,0 SCSI Multimedia CD-ROM Drive (650 MB)

Manufacturer................IBM Machine Type and Model......CDRM00203 ROS Level and ID............8B08 Device Specific.(Z0)........058002028F000018 Part Number.................73H2600 EC Level....................D75458A FRU Number..................73H2601 siota0 sa0 tty0 sa1 tty1 proc0 mem0 00-00-0T 00-00-S1 00-00-S1-00 00-00-S2 00-00-S2-00 00-00 00-0A Tablet Adapter Standard I/O Serial Port 1 Asynchronous Terminal Standard I/O Serial Port 2 Asynchronous Terminal Processor 32 MB Memory SIMM

Size........................32 Device Specific.(Z3)........90000000 EC Level....................00 mem1 00-0B 32 MB Memory SIMM

Size........................32 Device Specific.(Z3)........90000000 EC Level....................00 . . . ################################################# System Installed Filesets Fileset Level State Type Description (Uninstaller) ----------------------------------------------------------------------

Listing 13.3 AIXsysconfig.ksh shell script in action. (continued)

Taking a System Snapshot

363

. . . Tivoli_Management_Agent.client.rte 3.2.0.0 runtime” X11.Dt.ToolTalk 5.1.0.0 Support X11.Dt.adt 5.1.0.0 Developers’ X11.Dt.bitmaps X11.Dt.compat X11.Dt.helpinfo and Volumes X11.Dt.helpmin Files X11.Dt.helprun X11.Dt.lib Libraries X11.Dt.rte Environment X11.Dt.xdt2cde Tool X11.adt.bitmaps Application Bitmap Files X11.adt.imake Application imake X11.adt.include Application Include X11.adt.lib Application Libraries X11.adt.motif Application Motif X11.apps.xterm Application 5.1.0.0 5.1.0.0 5.1.0.0 5.1.0.0 5.1.0.0 5.1.0.0 5.1.0.0 5.1.0.0

C C C

F F F

Management Agent AIX CDE ToolTalk AIX CDE Application Toolkit AIX CDE Bitmaps AIX CDE Compatibility AIX CDE Help Files AIX CDE Minimum Help AIX CDE Runtime Help AIX CDE Runtime AIX Common Desktop (CDE) 1.0 AIX CDE Migration AIXwindows Development Toolkit

C C C C C C C

F F F F F F F

5.1.0.0 5.1.0.0

C C

F F

5.1.0.0

C

F

AIXwindows Development Toolkit

5.1.0.0

C

F

AIXwindows Development Toolkit Files AIXwindows Development Toolkit

C

F

5.1.0.0

C

F

AIXwindows Development Toolkit

5.1.0.0

C

F

AIXwindows xterm

Listing 13.3 AIXsysconfig.ksh shell script in action. (continues)

Free & Share & Open

364

Chapter 13

X11.base.common Common X11.base.lib Libraries X11.base.rte Environment bos.acct bos.adt.base Development bos.adt.debug Development bos.adt.include Development bos.adt.lib Development bos.adt.libm Development

5.1.0.0

C

F

AIXwindows Runtime Directories AIXwindows Runtime AIXwindows Runtime Accounting Services Base Application Toolkit Base Application Debuggers Base Application Include Files Base Application Libraries Base Application Math Library

5.1.0.0 5.1.0.1 5.1.0.0 5.1.0.1

C A C A

F F F F

5.1.0.1

A

F

5.1.0.1

A

F

5.1.0.0

C

F

5.1.0.0

C

F

bos.alt_disk_install.boot_images 5.1.0.0 Installation Disk bos.alt_disk_install.rte Installation 5.1.0.0

C

F

Alternate Disk Boot Images Alternate Disk Runtime Common Hardware Hardware Diagnostics Hardware Diagnostics

C

F

bos.diag.com 5.1.0.0 Diagnostics bos.diag.rte 5.1.0.0 bos.diag.util 5.1.0.1 Utilities bos.msg.en_US.net.tcp.client 5.1.0.0 U.S. English bos.msg.en_US.rte 5.1.0.0 Runtime bos.msg.en_US.svprint Subsystem 5.1.0.0

C C A

F F F

C C

F F

TCP/IP Messages Base Operating System Msgs - U.S. English System V Print Messages - U.S.

C

F

English bos.msg.en_US.sysmgt.nim.master_gui 4.3.0.0 U.S. English

C

F

NIM GUI Messages -

Listing 13.3 AIXsysconfig.ksh shell script in action. (continued)

Taking a System Snapshot

365

bos.msg.en_US.txt.tfs Services English bos.net.ate Emulator bos.net.ipsec.rte bos.net.ncs System 1.5.1 bos.net.nfs.adt Development bos.net.nfs.cachefs bos.net.nfs.client Client bos.net.nfs.server Server bos.net.nis.client Service . . . State A -B -C -O -? --

5.1.0.0

C

F

Text Formatting Messages - U.S.

5.1.0.0 5.1.0.0 5.1.0.0 5.1.0.0

C C C C

F F F F

Asynchronous Terminal IP Security Network Computing Network File System Toolkit CacheFS File System Network File System Network File System Network Information

5.1.0.0 5.1.0.1 5.1.0.0 5.1.0.0

C A C C

F F F F

codes: Applied. Broken. Committed. Obsolete. (partially migrated to newer version) Inconsistent State...Run lppchk -v.

Type codes: F -- Installp Fileset P -- Product C -- Component T -- Feature R -- RPM Package ################################################# Looking for Broken Filesets lppchk: The following filesets need to be installed or corrected to bring the system to a consistent state: vac.C.readme.ibm 4.4.0.1 (not installed; requisite fileset)

#################################################

Listing 13.3 AIXsysconfig.ksh shell script in action. (continues)

Free & Share & Open

366

Chapter 13

List of the last 100 users to log in to yogi root root root root root root root root shutdown root reboot reboot root root root . . . ftp ftp tty0 ftp pts/1 pts/0 pts/0 ftp tty0 ftp ~ ~ pts/3 pts/2 pts/1 booboo booboo booboo mrranger mrranger mrranger booboo booboo Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jul Jun Jun Jun Jun 25 25 24 24 24 24 24 23 10 09 09 27 26 26 26 13:28 12:17 17:35 17:35 17:11 17:11 17:09 21:53 00:25 23:41 19:38 16:07 20:55 20:55 20:55 - 13:29 (00:00) - 12:18 (00:00) still logged in. - 17:35 (00:00) still logged in. still logged in. - 17:11 (00:01) - 21:53 (00:00) - 23:41 (00:00)

mrranger mrranger mrranger

- 20:56 - 20:56 - 20:56

(00:00) (00:00) (00:00)

wtmp begins

Jul 31 18:20

This report is saved in: /usr/local/reboot/sys_snapshot.yogi.072502_094658

Listing 13.3 AIXsysconfig.ksh shell script in action. (continued)

From Listing 13.3 you can see that we collected a lot of information about the system configuration. This is just a sample of what you can collect, and I will leave the specifics of the information you gather up to you. For each function that you add or change, be sure to test the response. Sometimes you may be surprised that you do not see any output. Some of the command output shown in Listing 13.3 does not have any output because my little system does not have the hardware that the query is looking for. If you expect output and there is not any, try redirecting standard error to standard output by using the following syntax: command 2>&1

Many commands send information type output to standard error, specified by file descriptor 2, instead of standard output, specified by file descriptor 1. First try the command without this redirection.

Taking a System Snapshot

367

Other Options to Consider
There can always be improvements to any shell script. The shell script presented in this chapter is intended to be an example of the process of gathering system information. You always want to query the system for as much information as you can. Notice that I did not add any database or application configuration/statistics gathering here. The amount of information gathered is up to you. As I said before, every shop is different, but they are all the same when troubleshooting a problem. The AIXsysconfig.ksh shell script looks only at system-level statistics and configuration, so there is a large gap that you need to fill in. This gap is where your specific application comes into play. Look at your database and application documentation for the best method of gathering information about these products. By running the configuration gathering script at least once a week, you will save yourself a lot of effort when a problem arises.

Summary
In this chapter we strictly looked at AIX. The process is the same for any Unix flavor, but the information gathered will vary in each shop. No rocket science is needed here, but you do need a good understanding of how your system is configured. You need to understand the applications and databases and what determines a failed application. You may be looking for a set of processes, or it could be a database query with an SQL statement. These are the things that need research on your part to make this type of shell script really beneficial. In the next chapter we are going to move on to installing, configuring, and using sudo. The sudo program stands for super user do, and it allows us to set up specific commands that a user can execute as root. I hope you enjoyed this chapter, and I’ll see you in the next chapter!

Free & Share & Open

CHAPTER

14
Compiling, Installing, Configuring, and Using sudo

The main job of any good Systems Administrator is to protect the root password. No matter how firm and diligent we are about protecting the root password we always have the application support group and DBAs wanting root access for one reason or another. But, alas, there is a way to give specific users the ability to run selected commands as the root user without the need to know the root password. Facilitating this restricted root access is a free software program called sudo, which stands for superuser do. In this chapter we are going to show how to compile, install, configure, and use the sudo program. The current distribution can be downloaded by following the link on the Web site that accompanies the book, and I will list some Web other sites where you can download the program in this chapter. Because sudo is not a shell script you may be asking, “Why is sudo included in this book?” I am including the sudo chapter because I have not found any reference to sudo in any scripting book, and it is a nice tool to use. We will cover a short shell script at the end of this chapter showing how to use sudo in a shell script.

The Need for sudo
In Unix the root user is almighty and has absolutely no restrictions. All security is bypassed, and anyone with root access can perform any task, with some possibly resulting in major damage to the system, without any restrictions at all. Unix systems

369

Free & Share & Open

370

Chapter 14 do not ask “Are you sure?”; they just run the command specified by the root user and assume you know what you are doing. The sudo program allows the Systems Administrator to set up specific commands (or all commands) to be executed as the root user and specify only certain users (or groups of users) to execute the individual commands. In addition, all commands and command arguments are logged either to a defined file or the system syslog. The logging allows the Systems Administrator to have an audit trail and to monitor user sudo activities as well as failed sudo attempts! The user executes a restricted command by preceding the command with the word sudo. For example: sudo chmod 600 /etc/sudoers Password:

When a user executes the preceding command, a password prompt is displayed. The password that the sudo program is asking for is not the root password but the user’s password that wants root access. When the password is entered, the /etc/sudoers file is searched to determine if root authority should be granted to run the specified command. If both the system password is correct and the /etc/sudoers search grants access, then the command will execute with root authority. After this initial sudo command, the user may submit more sudo commands without the need for a password until a sudo timeout, typically five minutes without issuing another sudo command. After the timeout period the user will again be prompted for his or her password when a sudo command is entered.

Downloading and Compiling sudo
The sudo program is included on the Web site that accompanies the book and can be downloaded from various FTP mirror sites. The main sudo Web site is located at www.courtesan.com/sudo. The sudo program is free software and is distributed under a BDS-style license. As of this writing the current version of sudo is 1.6.6 and was released April 25, 2002. Todd Miller currently maintains the sudo program, and if you would like to tip Todd for his fine work you may do so at PayPal, which can be accessed from a link on the sudo main page. You can download sudo from any of the Web sites shown in Listing 14.1. http://www.courtesan.com/sudo/dist/ (Main site in Boulder, Colorado USA) http://www.rge.com/pub/admin/sudo/ (Rochester, New York USA) http://sudo.stikman.com/ (Los Angeles, California USA) http://www.c0r3dump.com/sudo/ (Edmonton, Canada) http://core.ring.gr.jp/archives/misc/sudo/ (Japan) http://www.ring.gr.jp/archives/misc/sudo/ (Japan) http://sudo.cdu.elektra.ru/ (Russia)

Listing 14.1 Web sites to download the sudo program.

Compiling, Installing, Configuring, and Using sudo
There are two ways to download the files. You can download the precompiled binaries for your Unix flavor and version or download the source code distribution and compile the sudo program for your particular machine. I always download the source code and compile it on each individual system. The process takes just a few minutes, and you can be assured that it will run on your system. If you have a boatload of systems to install, you may want to consider using the precompiled binaries and pushing the binaries out to each system, or writing a shell script to push and install the product! Either way you choose, you will need only about 4MB of free space to work with. Once sudo is installed you can remove the downloaded files if you need to regain the disk space. In this chapter we are going to download the source code and compile sudo for a particular system.

371

Compiling sudo
You will need a C compiler; cc is preferred but gcc normally works fine and is free to download. I say gcc normally works fine because I have found instances where gcc had compiler errors and cc did not have any problems. The source code distribution is in a compressed tar format, where gzip is used for compression. The gzip file has a .gz extension—for example, sudo-1.6.6.tar.gz. When you download the file, put the software distribution in a directory that has about 4MB of free space. In our example we will use /usr/local, which is a separate filesystem from /usr on my machine. You must have root access to compile, install, and configure sudo! After the sudo distribution file is placed in a work directory, the first step is to unzip the compressed file. The gunzip command uncompresses a gzipped file, as shown in the next example: gunzip sudo-1.6.6.tar.gz

After the file is uncompressed, you are left with the following tar archive file: sudo-1.6.6.tar When we untar the archive, a subdirectory will be created called sudo-1.6.6 that will contain all of the source code, LICENSE, README, manuals, configure, and Makefile. In the directory containing the sudo-1.6.6.tar file, in our case /usr/local, issue the following command: tar xvf sudo-1.6.6.tar

After the program distribution file is uncompressed and untarred we can proceed to the installation process. This is not a difficult process so if you have never worked with the make command and Makefile before, don’t worry. The first step is to configure the Makefile for your system. As you might expect, this is done with the configure command. First change directory to where the source code is located, in our example /usr/local/sudo-1.6.6, and run the configure command.

Free & Share & Open

372

Chapter 14 cd /usr/local/sudo-1.6.6 ./configure

The configure command goes through system checks and builds a Makefile and the config.h file used to build sudo for your system. The configure command output for my system is shown in Listing 14.2.
Configuring Sudo version 1.6.6 checking whether to lecture users the first time they run sudo... yes checking whether sudo should log via syslog or to a file by default... syslog checking which syslog facility sudo should log with... local2 checking at which syslog priority to log commands... notice checking at which syslog priority to log failures... badpri checking how long a line in the log file should be... 80 checking whether sudo should ignore ‘.’ or ‘’ in $PATH... no checking whether to send mail when a user is not in sudoers... yes checking whether to send mail when user listed but not for this host... no checking whether to send mail when a user tries a disallowed command... no checking who should get the mail that sudo sends... root checking for bad password prompt... Password: checking for bad password message... Sorry, try again. checking whether to expect fully qualified hosts in sudoers... no checking for umask programs should be run with... 0022 checking for default user to run commands as... root checking for editor that visudo should use... vi checking whether to obey EDITOR and VISUAL environment variables... no checking number of tries a user gets to enter their password... 3 checking time in minutes after which sudo will ask for a password again... 5 checking time in minutes after the password prompt will time out... 5 checking whether to use per-tty ticket files... no checking whether to include insults... no checking whether to override the user’s path... no checking whether to get ip addresses from the network interfaces... yes checking whether to do user authentication by default... yes checking whether to disable running the mailer as root... no checking whether to disable use of POSIX saved ids... no checking whether to disable shadow password support... no checking whether root should be allowed to use sudo... yes checking whether to log the hostname in the log file... no checking whether to invoke a shell if sudo is given no arguments... no checking whether to set $HOME to target user in shell mode... no checking whether to disable ‘command not found’ messages... no checking for egrep... egrep

Listing 14.2 Command output—./configure.

Compiling, Installing, Configuring, and Using sudo

373

checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking

for gcc... no for cc... cc for C compiler default output... a.out whether the C compiler works... yes whether we are cross compiling... no for executable suffix... for object suffix... o whether we are using the GNU C compiler... no whether cc accepts -g... (cached) no for POSIXized ISC... no for cc option to accept ANSI C... none needed how to run the C preprocessor... cc -E for uname... uname for tr... tr for sed... sed for nroff... nroff build system type... powerpc-ibm-aix5.1.0.0 host system type... powerpc-ibm-aix5.1.0.0 for getspnam... no for getspnam in -lgen... no for getprpwnam... no for an ANSI C-conforming const... yes for working volatile... yes for bison... no for byacc... no for mv... /usr/bin/mv for bourne shell... /bin/sh for sendmail... /usr/sbin/sendmail for vi... /usr/bin/vi for ANSI C header files... yes for dirent.h that defines DIR... yes for opendir in -ldir... no for malloc.h... yes for paths.h... yes for utime.h... yes for netgroup.h... no for sys/sockio.h... no for sys/bsdtypes.h... no for sys/select.h... yes POSIX termios... yes for sys/types.h... yes for sys/stat.h... yes for stdlib.h... yes for string.h... yes for memory.h... yes for strings.h... yes for inttypes.h... yes

Listing 14.2 Command output—./configure. (continues)

Free & Share & Open

374

Chapter 14

checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking checking

for stdint.h... no for unistd.h... yes for mode_t... yes for uid_t in sys/types.h... yes for sig_atomic_t... yes for sigaction_t... no for size_t... yes for ssize_t... yes for dev_t... yes for ino_t... yes for full void implementation... yes max length of uid_t... 10 for long long support... yes for sa_len field in struct sockaddr... yes return type of signal handlers... void for strchr... yes for strrchr... yes for memchr... yes for memcpy... yes for memset... yes for sysconf... yes for tzset... yes for seteuid... yes for setegid... yes for strftime... yes for setrlimit... yes for initgroups... yes for fstat... yes for setreuid... yes for getifaddrs... no for getcwd... yes for lockf... yes for waitpid... yes for innetgr... yes for getdomainname... yes for lsearch... yes for utime... yes for POSIX utime... yes for working fnmatch with FNM_CASEFOLD... no for isblank... yes for strerror... yes for strcasecmp... yes for sigaction... yes for snprintf... yes for vsnprintf... yes for asprintf... no for vasprintf... no

Listing 14.2 Command output—./configure. (continued)

Compiling, Installing, Configuring, and Using sudo

375

checking for crypt... yes checking for socket... yes checking for inet_addr... yes checking for syslog... yes checking for log file location... /var/adm/sudo.log checking for timestamp file location... /tmp/.odus configure: creating ./config.status config.status: creating Makefile config.status: creating sudo.man config.status: creating visudo.man config.status: creating sudoers.man config.status: creating config.h config.status: creating pathnames.h

Listing 14.2 Command output—./configure. (continued)

After the configure command completes without error, you have a customized Makefile for your system. You can, if you need to, edit the Makefile and change the default paths and the compiler to use. Now that we have a new customized Makefile we can now compile the sudo program on the system. Issue the following command, assuming you are still in the /usr/local/sudo-1.6.6 directory: make The make command is located in /usr/bin/make on most systems, and it uses the Makefile in the current directory to compile, in our case /usr/local/ sudo-1.6.6. The make command output is shown in Listing 14.3. Notice that my system uses the cc compiler.

cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 check.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 env.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 getspwuid.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 goodpath.c

Listing 14.3 Command output—make command. (continues)

Free & Share & Open

376

Chapter 14

cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 fileops.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 find_path.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 interfaces.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 logging.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 parse.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 set_perms.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 sudo.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 tgetpass.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 ./auth/sudo_auth.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 ./auth/passwd.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 sudo.tab.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 lex.yy.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 alloc.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 defaults.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 fnmatch.c cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 snprintf.c

Listing 14.3 Command output—make command. (continued)

Compiling, Installing, Configuring, and Using sudo

377

cc -o sudo check.o env.o getspwuid.o goodpath.o fileops.o find_path.o interfaces.o logging.o parse.o set_perms.o sudo.o tgetpass.o sudo_auth.o passwd.o sudo.tab.o lex.yy.o alloc.o defaults.o fnmatch.o snprintf.o -Wl,-bI:./aixcrypt.exp cc -c -I. -I. -D_XOPEN_EXTENDED_SOURCE D_PATH_SUDOERS=\”/etc/sudoers\” -D_PATH_SUDOERS_TMP=\”/etc/sudoers.tmp\” -DSUDOERS_UID=0 -DSUDOERS_GID=0 -DSUDOERS_MODE=0440 visudo.c cc -o visudo visudo.o fileops.o goodpath.o find_path.o sudo.tab.o lex.yy.o alloc.o defaults.o fnmatch.o snprintf.o Target “all” is up to date.

Listing 14.3 Command output—make command. (continued)

After the make command completes, we have custom compiled code for your system, but we still have one more installation step to complete before we are ready to configure sudo. This last step is to install the compiled files created with the make command. The next command handles the installation of sudo: make install

Remember that the make command is usually located in /usr/bin and should be in your $PATH. The output of the make install command for my machine is shown in Listing 14.4.
/bin/sh ./mkinstalldirs /usr/local/bin /usr/local/sbin /etc /usr/local/man/man8 /usr/local/man/man5 /bin/sh ./install-sh -c -O 0 -G 0 -M 4111 -s sudo /usr/local/bin/sudo /bin/sh ./install-sh -c -O 0 -G 0 -M 0111 -s visudo /usr/local/sbin/visudo test -f /etc/sudoers || /bin/sh ./install-sh -c -O 0 -G 0 -M 0440 ./sudoers /etc/sudoers /bin/sh ./install-sh -c -O 0 -G 0 -M 0444 ./sudo.man /usr/local/man/man8/sudo.8 /bin/sh ./install-sh -c -O 0 -G 0 -M 0444 ./visudo.man /usr/local/man/man8/visudo.8 /bin/sh ./install-sh -c -O 0 -G 0 -M 0444 ./sudoers.man /usr/local/man/man5/sudoers.5 Target “install” is up to date.

Listing 14.4 Command output—make install.

If you did not have any failures during the compilation and installation processes, then sudo is installed but not yet configured. In the next section we will look at two sample configuration files.

Free & Share & Open

378

Chapter 14

Configuring sudo
Configuring sudo is where a lot of people get a bit confused. The configuration is not too difficult if you take small steps and test each part as you build the configuration file. If you look in /etc after the installation is complete, you will see a file called sudoers. The sudoers file is used to configure the commands and users for the sudo program. Be very careful to never directly edit the sudoers file! A special program is supplied that has a wrapper around the vi editor called visudo, or vi sudo. The visudo program resides in /usr/local/sbin by default. The nice thing about visudo is that it checks the /etc/sudoers file for any errors before saving the file. If errors are detected, the visudo program will tell you exactly what the error is and in most cases the line the error is on. If you directly edit the /etc/sudoers file and you make a mistake, the editor will just let you save the file, with the mistake, and it can be difficult to find the error. The visudo program checks for the correct file format and ensures that the command/user references are consistent. If you make a mistake with a user name, the visudo editor will not catch the mistake, but this type of error should be easy to find and correct after an initial run. I am enclosing two samples of a /etc/sudoers file for you to use as a template in Listings 14.5 and 14.6.

N OT E The sudoers file in Listing 14.5 is used with the permission of Todd Miller at www.courtesan.com and is included in the sudo distribution as a sample. Thank you, Todd!

# # Sample /etc/sudoers file. # # This file MUST be edited with the ‘visudo’ command as root. # # See the sudoers man page for the details on how to write a sudoers file. # ## # User alias ## User_Alias User_Alias User_Alias

specification FULLTIMERS = millert, mikef, dowdy PARTTIMERS = bostley, jwfox, crawl WEBMASTERS = will, wendy, wim

## # Runas alias specification ##

Listing 14.5 Sample /etc/sudoers file #1.

Compiling, Installing, Configuring, and Using sudo

379

Runas_Alias Runas_Alias

OP = root, operator DB = oracle, sybase

## # Host alias specification ## Host_Alias SPARC = bigtime, eclipse, moet, anchor:\ SGI = grolsch, dandelion, black:\ ALPHA = widget, thalamus, foobar:\ HPPA = boa, nag, python Host_Alias CUNETS = 128.138.0.0/255.255.0.0 Host_Alias CSNETS = 128.138.243.0, 128.138.204.0/24, 128.138.242.0 Host_Alias SERVERS = master, mail, www, ns Host_Alias CDROM = orion, perseus, hercules ## # Cmnd alias specification ## Cmnd_Alias DUMPS = /usr/sbin/dump, /usr/sbin/rdump, /usr/sbin/restore, \ /usr/sbin/rrestore, /usr/bin/mt Cmnd_Alias KILL = /usr/bin/kill Cmnd_Alias PRINTING = /usr/sbin/lpc, /usr/bin/lprm Cmnd_Alias SHUTDOWN = /usr/sbin/shutdown Cmnd_Alias HALT = /usr/sbin/halt, /usr/sbin/fasthalt Cmnd_Alias REBOOT = /usr/sbin/reboot, /usr/sbin/fastboot Cmnd_Alias SHELLS = /usr/bin/sh, /usr/bin/csh, /usr/bin/ksh, \ /usr/local/bin/tcsh, /usr/bin/rsh, \ /usr/local/bin/zsh Cmnd_Alias SU = /usr/bin/su Cmnd_Alias VIPW = /usr/sbin/vipw, /usr/bin/passwd, /usr/bin/chsh, \ /usr/bin/chfn ## # Override builtin defaults ## Defaults syslog=auth Defaults:FULLTIMERS !lecture Defaults:millert !authenticate Defaults@SERVERS log_year, logfile=/var/log/sudo.log ## # User specification ## # root and users in group wheel can run anything on any machine # as any user

Listing 14.5 Sample /etc/sudoers file #1. (continues)

Free & Share & Open

380

Chapter 14

root %wheel

ALL = (ALL) ALL ALL = (ALL) ALL

# full time sysadmins can run anything on any machine without a password FULLTIMERS ALL = NOPASSWD: ALL # part time sysadmins may run anything but need a password PARTTIMERS ALL = ALL # jack may run anything on machines in CSNETS jack CSNETS = ALL # lisa may run any command on any host in CUNETS (a class B network) lisa CUNETS = ALL # operator may run maintenance commands and anything in /usr/oper/bin/ operator ALL = DUMPS, KILL, PRINTING, SHUTDOWN, HALT, REBOOT,\ /usr/oper/bin/ # joe may su only to operator joe ALL = /usr/bin/su operator # pete may change passwords for anyone but root on the hp snakes pete HPPA = /usr/bin/passwd [A-z]*, !/usr/bin/passwd root # bob may run anything on the sparc and sgi machines as any user # listed in the Runas_Alias “OP” (ie: root and operator) bob SPARC = (OP) ALL : SGI = (OP) ALL # jim may run anything on machines in the biglab netgroup jim +biglab = ALL # users in the secretaries netgroup need to help manage the printers # as well as add and remove users +secretaries ALL = PRINTING, /usr/bin/adduser, /usr/bin/rmuser # fred can run commands as oracle or sybase without a password fred ALL = (DB) NOPASSWD: ALL # on the alphas, john may su to anyone but root and flags are not allowed john ALPHA = /usr/bin/su [!-]*, !/usr/bin/su *root* # jen can run anything on all machines except the ones # in the “SERVERS” Host_Alias

Listing 14.5 Sample /etc/sudoers file #1. (continued)

Compiling, Installing, Configuring, and Using sudo

381

jen

ALL, !SERVERS = ALL

# jill can run any commands in the directory /usr/bin/, except for # those in the SU and SHELLS aliases. jill SERVERS = /usr/bin/, !SU, !SHELLS # steve can run any command in the directory /usr/local/op_commands/ # as user operator. steve CSNETS = (operator) /usr/local/op_commands/ # matt needs to be able to kill things on his workstation when # they get hung. matt valkyrie = KILL # users in the WEBMASTERS User_Alias (will, wendy, and wim) # may run any command as user www (which owns the web pages) # or simply su to www. WEBMASTERS www = (www) ALL, (root) /usr/bin/su www # anyone can mount/unmount a CD-ROM on the machines in the CDROM alias ALL CDROM = NOPASSWD: /sbin/umount /CDROM,\ /sbin/mount -o nosuid\,nodev /dev/cd0a /CDROM

Listing 14.5 Sample /etc/sudoers file #1. (continued)

# sudoers file. # # This file MUST be edited with the ‘visudo’ command as root. # # See the sudoers man page for the details on how to write a sudoers file. # # Users Identification: # # All ROOT access: # # d7742 - Michael # # Restricted Access to: mount umount and exportfs # #

Listing 14.6 Sample /etc/sudoers file #2. (continues)

Free & Share & Open

382

Chapter 14

# # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #

Restricted Access to: Start and stop Fasttrack Web Server d3920 - Park d7525 - Brinker d7794 - Doan Restricted OPERATIONS access d6331 d6814 d8422 d9226 d9443 d0640 d2105 d2188 d3408 d3551 d3883 d6290 d2749 d6635 d3916 d6782 d6810 d6811 d6817 d5123 d7504 d7505 d7723 Sutter Martin Smith Milando Summers Lawson Fanchin Grizzle Foster Dennis Nations Alexander Mayo Wright Chatman Scott Duckery Wells Gilliam Crynick Davis McCaskey Rivers

Host alias specification LOCAL=yogi

Host_Alias

# User alias specification User_Alias NORMAL=d7742,d7537,d7526,d6029,d7204,d1076,d7764,d7808 User_Alias ADMIN=e17742,d7211,d6895,d8665,d7539,b003 User_Alias ORACLE=d7742 User_Alias SAP=d7742 User_Alias OPERATOR=d7742,d6895,d6331,d6814,d8422,d9226,d9443,d0640, d2105,d2188,d3408,d3551,d3883,d6290,d2749,d6635,d3916,d6782,d6810,

Listing 14.6 Sample /etc/sudoers file #2. (continued)

Compiling, Installing, Configuring, and Using sudo

383

d6811,d6817,d5123,d7504,d7505,d7723 User_Alias FASTTRACK=d3920,d7525,d7794 # Cmnd alias specification Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias MNT=/usr/bin/mount UMNT=/usr/bin/umount EXP_FS=/usr/bin/exportfs KILL=/usr/bin/kill ROOT_SU=/usr/bin/su SU_ROOT=/usr/bin/su - root SUROOT=/usr/bin/su root ORACLE_SU=/usr/bin/su - oracle SAP_SU=/usr/bin/su - sap TCPDUMP=/usr/sbin/tcpdump ERRPT=/usr/bin/errpt SVRMGRL=/oracle/product/8.0.5/bin/svrmgrl RSH_UPDATE=/usr/local/bin/rsh_update.ksh START_FT_YOGI=/usr/netscape/httpd-yogi/start STOP_FT_YOGI=/usr/netscape/httpd-yogi/stop START_FT_DINO=/usr/netscape/httpd-dino/start STOP_FT_DINO=/usr/netscape/httpd-dino/stop START_WSADM=/usr/netscape/start-admin STOP_WSADM=/usr/netscape/stop-admin

# User privilege specification # FULL ROOT ACCESS!!!!!! (BE CAREFUL GRANTING FULL ROOT!!!!!!!) root ALL=(ALL) ALL d7742 ALL=(ALL) ALL # Michael # Only mount, umount and exportfs NORMAL LOCAL=MNT,UMNT,EXP_FS # Some Limited Sys Admin Functions ADMIN LOCAL=MNT,UMNT,KILL,ORACLE_SU,SAP_SU,TCPDUMP,ERRPT,ROOT_SU: \ LOCAL=SU_ROOT,SUROOT,EXP_FS # Some Operator Functions OPERATOR LOCAL=RSH_UPDATE # Some FastTrack/WebAdm Functions FASTTRACK LOCAL=START_FT_E1,STOP_FT_E1,START_FT_E2,STOP_FT_E2,START_WSADM,

Listing 14.6 Sample /etc/sudoers file #2. (continues)

Free & Share & Open

384

Chapter 14

STOP_WSADM # Override Defaults # Change the default location of the SUDO log file Defaults logfile=/var/adm/sudo.log

Listing 14.6 Sample /etc/sudoers file #2. (continued)

As you can see by the two sample /etc/sudoers files, you can get as detailed as you want. As you look at these files, notice that there are four kinds of aliases: User_Alias, Runas_Alias, Host_Alias, and Cmd_Alias. The use of each alias type is listed next. A User_Alias is a list that can contain any combination of usernames, UID (with a “#” prefix), system groups (with a “%” prefix), netgroups (with a “+” prefix), and other user-defined aliases. Any of these can be prefixed with the NOT operator, “!”, to negate the entry. A Runas_Alias can contain any of the same elements as the User_Alias; the only difference is that you use Runas_Alias instead of User_Alias in the configuration. The Runas_Alias allows execution of a command as a user other than root. A Host_Alias is a list of hostnames, IP addresses, or netgroups (with a “+” prefix). The Host_Alias also supports the NOT operator, “!”, to negate an entry. You will need to use the fully qualified DNS name if the hostname command on any machine returns the name of the machine in a fully qualified DNS format. The visudo editor will not catch this “error.” A Cmnd_Alias is list of one or more commands specified by a full pathname, not just the filename. You can also specify directories and other aliases to commands. The command alone will allow command arguments to the command, but you can disable command arguments using double quotes (“ “). If a directory is specified a user can execute any command within that directory, but not any subdirectories. Wildcards are allowed, but be very careful to ensure that the wildcard is working as expected. I am not going to discuss every piece of sudo because very detailed documentation is included with the sudo distribution, and I need to limit my page count in this book. Our next step is to look at how to use sudo and how to use sudo in a shell script.

Using sudo
We use sudo by preceding the command that we want to run with the word sudo. As an example, if my user ID is rmichael and I want to gain root access for the first time, I will follow these steps:

Compiling, Installing, Configuring, and Using sudo

385

PATH=$PATH:/usr/local/bin export PATH sudo su - root We trust you have received the usual lecture from the local System Administrator. It usually boils down to these two things: #1) Respect the privacy of others. #2) Think before you type. Password: yogi@/#

Listing 14.7 Using sudo for the first time.

Notice the short lecture that is displayed in Listing 14.7. This lecture message is displayed only the first time that sudo is used by each user. In the password field the user responds with his or her normal user account password, not the root password. You should be careful granting full root permission like this. Allowing a user to su to root via the sudo program does not leave an audit trail of what the user did as root! You should still have the root history file if the user did not delete or edit the file. Also notice in Listing 14.7 that I added /usr/local/bin to my $PATH. By default the sudo command is located in the /usr/local/bin directory, but most shops do not add this directory to the $PATH environment variable as a normal path when setting up user accounts. Just make sure that all sudo users have the sudo command in the $PATH or that they need to provide the full pathname to the sudo command.

Using sudo in a Shell Script
We can also use sudo in a shell script. As you create the shell script add the sudo command as a prefix to each command that you want to execute as root. The script in Listing 14.8 uses the sudo command to allow our Operations Team to reset passwords. On an AIX system you can manage user passwords with the pwdadm command. In this particular shell script we want our Operations Team to be able to change a user’s password from a menu selection in a shell script. The bold text shown in Listing 14.8 points out the use of sudo and also the use of the tput command for reverse video, which we will study further in Chapter 15.

Free & Share & Open

386

Chapter 14

#!/usr/bin/ksh # # SCRIPT: chpwd_menu.ksh # AUTHOR: Randy Michael # DATE: 11/05/2001 # PLATFORM: AIX # REV: 1.1.P # # PURPOSE: This script was created for the Operations Team # to change user passwords. This shell script uses # “sudo” to execute the “pwdadm” command as root. # Each member of the Operations Team needs to be # added to the /etc/sudoers file. CAUTION: When # editing the /etc/sudoers file always use the # /usr/local/sbin/visudo program editor!!! # NEVER DIRECTLY EDIT THE sudoers FILE!!!! # # REV LIST: # # set -x # Uncomment to debug this script # set -n # Uncomment to check syntax without any execution # ####################################################### # DEFINE FUNCTIONS HERE ####################################################### function chg_pwd { USER_NAME=”$1” echo “\nThe next password prompt is for YOUR NORMAL PASSWORD” echo “NOT the new password...” # The next command turns off the checking of the password history /usr/local/bin/sudo /usr/bin/pwdadm -f NOCHECK $USER_NAME if [ $? -ne 0 ] then echo “\nERROR: Turning off password history failed...” usage exit 1 fi # The next command changes the user’s password /usr/local/bin/sudo /usr/bin/pwdadm $USER_NAME

Listing 14.8 chpwd_menu.ksh shell script listing.

Compiling, Installing, Configuring, and Using sudo

387

if [ $? -ne 0 ] then echo “\nERROR: Changing $USER_NAME password failed...” usage exit 1 fi # The next command forces the user to change his or her password # at the next login. /usr/local/bin/sudo /usr/bin/pwdadm -f ADMCHG $USER_NAME return 0 } ####################################################### # START OF MAIN ####################################################### OPT=0 clear # Initialize to zero # Clear the screen

while [[ $OPT != 99 ]] # Start a loop do # Draw reverse image bar across the top of the screen # with the system name. clear tput smso echo “ tput sgr0 echo “” # Draw menu options. echo “\n\n\n\n\n\n\n” print “10. Change Password” echo “\n\n\n\n\n\n\n\n\n” print “99. Exit Menu” # Draw reverse image bar across bottom of screen,

$(hostname)

“

Listing 14.8 chpwd_menu.ksh shell script listing. (continues)

Free & Share & Open

388

Chapter 14

# with error message, if any. tput smso echo “ tput sgr0

$MSG

“

# Prompt for menu option. read OPT # Assume invalid selection was taken. Message is always # displayed, so blank it out when a valid option is selected. MSG=” Invalid option selected # Option 10 - Change Password “

if [ $OPT -eq 10 ] then echo “\nUsername for password change? \c” read USERNAME grep $USERNAME /etc/passwd >/dev/null 2>&1 if [ $? -eq 0 ] then chg_pwd $USERNAME if [ $? -eq 0 ] then MSG=”$USERNAME password successfully changed” else MSG=”ERROR: $USERNAME password change failed” fi else MSG=” ERROR: Invalid username $USERNAME “ fi fi # End of Option 99 Loop done # Erase menu from screen upon exiting. clear

Listing 14.8 chpwd_menu.ksh shell script listing. (continued)

The chpwd_menu.ksh shell script in Listing 14.8 displays a menu on the screen that has only two options, change a user’s password or exit. This shell script uses the sudo program to execute the pwdadm command as the root user. The pwdadm command is

Compiling, Installing, Configuring, and Using sudo used for password administration in AIX and has options to turn password history checking off and to force password changes on the next login attempt. The pwdadm command is executed three times in the chg_pwd function within the shell script. The first time pwdadm is executed as root we turn off the checking of the password history. Notice that I added a comment to the staff that the next password prompt is for their normal user password, not the new user password. I turn off the history checking because the password that the Operation Team is going to enter is a temporary password. The next time the user logs in, the system will prompt for a new password, and at this stage the password history will be checked. The second time that pwdadm is executed, the password is actually changed by the Operations Team member. The third time pwdadm is executed, the user is forced to change his or her password the next time they log in. Each time sudo is used to execute pwdadm as root. Also notice the tput commands. The tput command has many options to control the cursor and the terminal. In this script we are using reverse video to display the hostname of the machine in the menu title bar at the top and to display messages at the bottom of the menu. There is much more on the tput command options in Chapter 15.

389

The sudo Log File
Before we end this chapter I want to show you what the sudo log file looks like. Each time that sudo is executed, an entry is made in the specified log. Logging can be to a file or to the system syslog. I specify a log file in the /etc/sudoers, but you may prefer the syslog. A short version of my sudo log file is shown in Listing 14.9.
Nov Nov Nov Jul Oct Nov Nov ; 9 10:07:44 : d7742 : TTY=pts/2 ; PWD=/usr/local ; USER=root ; COMMAND=/usr/bin/ftp bambam 9 10:09:13 : d7742 : TTY=pts/2 ; PWD=/usr/local ; USER=root ; COMMAND=/usr/bin/ftp dino 13 10:10:48 : d7742 : TTY=pts/0 ; PWD=/home/guest ; USER=root ; COMMAND=/usr/bin/whoami 23 17:35:47 : d7996 : TTY=pts/3 ; PWD=/home/guest ; USER=root ; COMMAND=/usr/sbin/mount /usr/local/common 2 09:29:33 : d7742 : TTY=pts/1 ; PWD=/home/d7742 ; USER=root ; COMMAND=/usr/bin/su 14 16:01:31 : d7742 : TTY=pts/0 ; PWD=/home/d7742 ; USER=root ; COMMAND=/usr/bin/su - root 14 16:03:58 : rmichael : TTY=pts/0 ; PWD=/home/rmichael ; USER=root

COMMAND=/usr/bin/su - root Nov 15 11:31:32 : d7742 : TTY=pts/0 ; PWD=/scripts ; USER=root ; COMMAND=/usr/bin/pwdadm -f NOCHECK rmichael Nov 15 11:31:32 : d7742 : TTY=pts/0 ; PWD=/scripts ; USER=root ; COMMAND=/usr/bin/pwdadm rmichael Nov 15 11:31:32 : d7742 : TTY=pts/0 ; PWD=/scripts ; USER=root ;

Listing 14.9 Sample sudo log file. (continues)

Free & Share & Open

390

Chapter 14

COMMAND=/usr/bin/pwdadm -f ADMCHG rmichael Nov 15 14:58:49 : root : TTY=pts/0 ; PWD=/usr/local/sudo-1.6.3p7 ; USER=root ; COMMAND=/usr/bin/errpt Nov 15 14:59:50 : d7742 : 3 incorrect password attempts ; TTY=pts/0 ; PWD=/home/d7742 ; USER=root ; COMMAND=/usr/bin/errpt

Listing 14.9 Sample sudo log file. (continued)

In Listing 14.9 notice the last line of output. This line shows three incorrect password attempts. You can set up sudo to send an email on each password failure if you want immediate notification of misuse. The shell script in Listing 14.8 produced three log entries on each password change. I have highlighted several other entries for ftp and su to root to show you how the log entries look.

Summary
Through this chapter we have shown how to compile, install, configure, and use the sudo program. We all know that protecting the root password is one of our main tasks as a Systems Administrator, and sudo makes the job a little less difficult. When you use sudo in a shell script, it is important that each user is familiar with sudo and has used it at least once from the command line. Remember that on the first use the lecture message is displayed and you do not want a lecture in the middle of a menu! In the sudo distribution there are several files that you should review. The README file has valuable information in installation and a lot of OS-specific problems and workarounds. The FAQ file answers the most frequently asked questions. The Sudoers Manual is a must read! This manual describes the many options in configuring your sudoers file. Finally, we have the Visudo Manual that explains how to use the visudo editor and lists the command options and possible error conditions. Again, I want to thank Todd Miller at www.courtesan.com for allowing me to use his material in this chapter. In the next chapter we are going to create a highlight grep script. If you have ever wanted to find a text string in a large file, you will really appreciate this script! The command syntax is exactly the same as the grep command, but instead of extracting the line that the grep command pattern matched on, we display the entire file and use reverse video to highlight the text within the file.

CHAPTER

15 hgrep: Highlighted grep Script

Ever want to find text in a large file easily? The larger the text file, the more you will appreciate this shell script. We can use reverse video in shell scripts for more than just making pretty menus. What about highlighting text in a file or in a command’s output? In this chapter we are going to show an example of using reverse video in a shell script that works similar to the grep command. Instead of displaying the line(s) that match the pattern, we are going to display the entire file, or command output, with the matched pattern highlighted in reverse video. I like to call this hgrep. In the process of creating this shell script, an initial test script was developed that ended up being very complicated. It started by grepping each line for the specified pattern. If the pattern was found in the line, then a scan of the line, character by character, was started to locate the exact pattern in the line for highlighting, then we grepped again for the pattern in remaining line of text, and so on. This initial code had quite a few problems, other than the complicated nature of the script, caused by Unix special characters making the output do some very interesting things when scanning shell script code. Regular text files worked fine, but the script was very slow to execute. Then there was the revelation that sed should somehow be able to handle the pattern matching—and do so a lot faster than parsing the file with a shell script. A Korn shell script is really not meant to work on a file line by line and character by character; it can be done, but this is what Perl is for! The problem to resolve using sed was how to add in the highlighting control within a sed command statement. After thinking about using sed and command substitution for a while, I had a working script in about 15 minutes (we might have a record!), and the following is what I came up with.

391

Free & Share & Open

392

Chapter 15

Reverse Video Control
There are two commands that control reverse video: tput smso turns soft reverse video on, and tput rmso turns highlighting back off. The tput command has many other options to control the terminal, but tput sgr0 (sgr-zero) will turn every tput option off. To highlight text we turn reverse video on, print whatever we want highlighted, and then turn reverse video off. We can also save this output, with the highlighted text, in a file. To display the file with highlighted text we can use pg, or page, and on some operating systems more will work. The more command did not work on either AIX or HP-UX operating systems. Instead, the more command displayed the characters that make up the escape sequence for the highlighted text, not the highlighted text itself. You would see the same result using the vi editor. On Solaris both commands displayed the highlighted text, but not all operating systems have the pg and page commands. There is one common mistake that will prevent this shell script from working, not double quoting the variables, for example “$STRING”. The double quotes have no effect on a single-word pattern match, but for multiword string patterns the variables must be double quoted or standard error will produce command usage errors within the script. The errors are due to the fact that each word that makes up the string pattern will be interpreted as a separate argument instead of one entity. The double quotes are very important when working with string variables. Forgetting the double quotes is a very hard error to find when troubleshooting code! The sed command is next. Remember the basic sed syntax that we use in this book: cat $FILENAME | sed s/current_string/new_string/g

In our script we want to take the sed command statement and redirect output to a file, then display the file with pg, page, or more. cat $FILENAME | sed s/current_string/new_string/g > $OUTPUT_FILE pg $OUTPUT_FILE --verse-more $OUTPUT_FILE

To add in the reverse video piece we have to do some command substitution within the sed statement using the tput commands—this is the part that had to be worked out. Where we specify the new_string we will add in the control for reverse video using command substitution, one to turn highlighting on and one to turn it back off. When the command substitution is added, our sed statement will look like the following: sed s/current_string/$(tput smso)new_string$(tput rmso)/g

In our case, the current_string and new_string will be the same because we only want to highlight existing text without changing it. We also want the string to be assigned to a variable, as in the next command: sed s/”$STRING”/$(tput smso)”$STRING”$(tput rmso)/g

hgrep: Highlighted grep Script
Notice the double quotes around the string variable, “$STRING”. Do not forget to add the double quotes around variables! As an experiment using command substitution, try this next command statement on any Unix machine: cat /etc/hosts | sed s/`hostname`/$(tput smso)`hostname`$(tput rmso)/g

393

In the preceding command statement notice that we used both types of command substitution, enclosing the command within back tics, `command`, and the dollar parentheses method, $(command). The previous statement will cat the /etc/hosts file and highlight the machine’s hostname in reverse video each time it appears in the file. Now try the same command, but this time pipe the command to more. Try the same command again using pg and page instead of more, if your machine supports the page commands. If your machine does not have the pg command, then the more command should work. If your operating system has both pg and more, notice that using more may not display the string pattern in reverse video—it will display the characters that make up the escape sequence that the tput commands create, but Solaris is an exception. We will need to consider this when we display the result on different operating systems. To make this script have the same look and feel as the grep command, we want to be able to supply input via a file, as a command-line argument, or as standard input from a command pipe. When supplying a filename to the script as a command-line argument, we need to ensure that the file exists, its size is greater than zero bytes, it is readable by this script, and the string pattern is matched in the file. We could leave out the last step, but if the pattern is not in the file then it would be nice to let the user know. If we are getting input from standard input instead of a file specified as an argument— for example, cat /etc/hosts | hgrep.ksh `hostname`—then we need to check for the string pattern in the output file instead of the input file. Then we can still inform the user if the pattern is not found.

Building the hgrep.ksh Shell Script
Now that we have the basic command syntax, let’s build the hgrep.ksh shell script. There are two types of input for this script, file input and standard input. For the file input we need to do some sanity checks so that we don’t get standard error messages from the system. We also want to give the user some feedback if there is something that will cause an error using the specified file as input—for example, the file does not exist or is not readable by the script because of file permissions. The command syntax using the hgrep.ksh script should be the same as the grep command, which is: grep pattern [filename]

By looking at this we can determine that we will sanity-check the file only when we have two command-line arguments; otherwise, we are using piped-in standard input, which implies that we check the file only when $# is equal to 2. We begin with checking the command-line arguments and making assignments of the arguments to variables.

Free & Share & Open

394

Chapter 15 if [ $# -eq 1 ] then # Input coming from standard input PATTERN=”$1” # Pattern to highlight FILENAME= # Assign NULL to FILENAME elif [ $# -eq 2 ] then # Input coming from $FILENAME file PATTERN=”$1” # Pattern to highlight FILENAME=”$2” # File to use as input else # Incorrect number of command-line arguments usage exit 1 fi

We should now have enough to get us started. If we have a single command-line argument, then we assign $1 to PATTERN and assign the FILENAME variable a NULL value. If there are two command-line arguments, then we assign $1 to PATTERN and $2 to FILENAME. If we have zero or more than two arguments, then we display the usage message and exit with a return code of 1, one. The function for correct usage is listed here: function usage { echo “\nUSAGE: $SCRIPT_NAME }

pattern

[filename]\n”

Follow through the hgrep.ksh script in Listing 15.1, and the process will be explained at the end of the shell script.

#!/usr/bin/ksh # # SCRIPT: hgrep.ksh # AUTHOR: Randy Michael # DATE: 03/09/2001 # REV 2.1.P # # PLATFORM: Not Platform Dependent...(Not very platform dependent) # There is a slight “more” command issue that has been # resolved # # PURPOSE: This script is used to highlight text in a file or standard

Listing 15.1 hgrep.ksh shell script.

hgrep: Highlighted grep Script

395

# # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # # #

input. Given a text string and a file, or standard input, the script will search for the specified string and highlight each occurrence of the string using command substitution within a sed statement to turn on and off the reverse video. “tput smso” turns on reverse video and “tput rmso” will turn it off. This script is a “highlighted grep” command. set -x set -n EXIT CODES: 0 1 2 3 ==> ==> ==> ==> Script exited normally Usage error Input file error Pattern not found in the file # Uncomment to debug # Uncomment to check command syntax without execution

REV LIST: 03/12/2001 - Randy Michael - Sr. Sys. Admin. Added code to just exit if the string is not in the target file. 03/13/2001 - Randy Michael - Sr. Sys. Admin. Added code to ensure the target file is a readable “regular” non-zero file. 03/13/2001 - Randy Michael - Sr. Sys. Admin. Added code to highlight the text string and filename in the error and information messages. 08-22-2001 - Randy Michael - Sr. Sys. Admin Changed the code to allow this script to accept standard input from a pipe. This makes the script work more like the grep command

SCRIPT_NAME=`basename $0` ############################################## ########### DEFINE FUNCTIONS HERE ############ ############################################## function usage { echo “\nUSAGE: $SCRIPT_NAME }

pattern

[filename]\n”

############################################## ########### CHECK COMMAND SYNTAX #############

Listing 15.1 hgrep.ksh shell script. (continues)

Free & Share & Open

396

Chapter 15

############################################## if [ $# -eq 1 ] then # Input coming from standard input PATTERN=”$1” # Pattern to highlight FILENAME= # Assign NULL to FILENAME elif [ $# -eq 2 ] then # Input coming from $FILENAME file PATTERN=”$1” # Pattern to highlight FILENAME=”$2” # File to use as input # Perform sanity checks on the file!!! # Does the file exist as a “regular” file? if [[ ! -f $FILENAME ]] then echo “\nERROR: \c” tput smso echo “${FILENAME}\c” # Highlight the filename tput rmso echo “ does not exist as a regular file...\n” usage exit 2 fi # Is the file empty? if [[ ! -s $FILENAME ]] then echo “\nERROR: \c” tput smso echo “${FILENAME}\c” # Highlight the filename tput rmso echo “ file size is zero...nothing to search\n” usage exit 2 fi # Is the file readable by this script? if [[ ! -r $FILENAME ]] then

Listing 15.1 hgrep.ksh shell script. (continued)

hgrep: Highlighted grep Script

397

echo “\nERROR: \c” tput smso echo “${FILENAME}\c” # Highlight the filename tput rmso echo “ is not readable to this program...\n” usage exit 2 fi # Is the pattern anywhere in the file? grep “$PATTERN” $FILENAME >/dev/null 2>&1 if [ $? -ne 0 ] then echo “\nSORRY: The string \c” tput smso echo “${PATTERN}\c” # Highlight the pattern tput rmso echo “ was not found in \c” tput smso echo “${FILENAME}\c” # Highlight the filename tput rmso echo “\n\n....EXITING...\n” exit 3 fi else # Incorrect number of command line arguments usage exit 1 fi ############################################## ########### DEFINE VARIABLES HERE ############ ############################################## OUTPUT_FILE=”/tmp/highlightfile.out” >$OUTPUT_FILE ############################################## ############ START OF MAIN ################### ############################################## # If the $FILENAME varaible is NULL then input is from a command pipe # Testing for NULL assigned to $FILENAME. if [[ ! -z “$FILENAME” && “$FILENAME” != ‘’ ]] then

Listing 15.1 hgrep.ksh shell script. (continues)

Free & Share & Open

398

Chapter 15

# Using $FILENAME as input

# MUST USE DOUBLE QUOTES AROUND $PATTERN!!! -> “$PATTERN” cat “$FILENAME” \ | sed s/”${PATTERN}”/$(tput smso)”${PATTERN}”$(tput rmso)/g \ > $OUTPUT_FILE else # Input is from standard input... # MUST USE DOUBLE QUOTES AROUND $PATTERN!!! -> “$PATTERN” sed s/”${PATTERN}”/$(tput smso)”${PATTERN}”$(tput rmso)/g \ > $OUTPUT_FILE # Check to see if the pattern was in the standard input grep “$PATTERN” $OUTOUT_FILE >/dev/null 2>&1 if [ $? -ne 0 ] then echo “\nSORRY: The string \c” tput smso echo “${PATTERN}\c” tput rmso echo “ was not found in standard input \c” echo “\n\n....EXITING...\n” exit 3 fi fi # # # # # Check the operating system, on AIX and HP-UX we need to use the “pg”, or “page” command. The “more” command does not work to highlight the text, it will show only the characters that make up the escape sequence. All other operating systems use the “more” command.

case $(uname) in AIX|HP-UX) # This is a fancy “pg” command. It acts similarly to the # “more” command but instead of showing the percentage # displayed it shows the page number of the file /usr/bin/cat $OUTPUT_FILE | /usr/bin/pg -csn -p”Page %d:” ;; *)

Listing 15.1 hgrep.ksh shell script. (continued)

hgrep: Highlighted grep Script

399

/usr/bin/cat $OUTPUT_FILE | /usr/bin/more ;; esac rm -f $OUTPUT_FILE # End of Script Cleanup

Listing 15.1 hgrep.ksh shell script. (continued)

In the shell script in Listing 15.1 we first check for the correct number of commandline arguments; either one or two arguments are valid. Otherwise, the script usage message is displayed, and the script will exit with a return code 1. If we have the correct number of arguments, then we assign the arguments to variables. If we have two command-line arguments, then an input file is specified in $2—at least it is supposed to be a file. We need to do some sanity checking on this second command-line argument by first checking to see that the file exists as a regular file. We do not want to do anything with the file if it is a block or character special file, a directory, or any other nonregular file. Next we make sure that the file is not empty. Then we ensure that the script can read the file, and finally we grep for the pattern in the file to see if we have anything to highlight. If all of the tests are passed, then we can proceed. By checking if the $FILENAME variable is null, or empty, we know which type of input we are dealing with. A null or empty $FILENAME variable means we use standard input, which is input from a pipe in this case. If $FILENAME is not null, then we have a file specified as input to the script on the command line. The only difference in handling an input file versus standard input is that we will supply the ”cat $FILENAME |” if there is an input file specified. Otherwise, the input is already coming in from a pipe directly into the sed statement—it’s that simple. We have one more check before displaying the output. If we are using piped-in standard input, then we grep for “$PATTERN” in the $FILENAME to see if it exists. If not, we display a string not found message and exit. The output display is interesting because more will not work on HP-UX or AIX to display the highlighted text. For HP-UX and AIX we use pg instead of more. To determine which flavor of Unix we are running, we use the uname command in a case statement. If the OS is either AIX or HP-UX, we used a fancy pg command, which has output that appears similar to the more output. Using pg -csn -p"Page %d:" will display the page number of the file, where more displays the percentage of file. All other Unix flavors will use more to display the output file. The script in Listing 15.1 is a good example of how a little ingenuity can greatly simplify a challenge. We sometimes make things more complicated than they need to be, as in my initial test script that parsed through the file line by line and character by character, searching for the pattern. We live and learn!

Free & Share & Open

400

Chapter 15

Other Options to Consider
As with every script there is room for improvement or customization, however you want to look at it.

Other Options for the tput Command
The only tput command option that we worked with was the tput smso command, which is used to turn on highlighting. The tput command has many other options to control terminal display. In our example we did a highlight of not only the text but also the surrounding block for each character. We could also highlight only the text piece, double video the entire text block, underline with other options—for example, we could have underlined bold text. The tput command is fun to play with. The short list of command options is shown in Table 15.1.

Table 15.1 tput bell tput blink tput bold tput civis

Options for the tput Command Ring the bell Start blinking mode Start double intensity (much brighter than reverse video) Turn the cursor off (make the cursor invisible) Make the cursor normal again Send a carriage to the terminal Make the cursor very bright Start one-half intensity mode Clear to the end of the display Clear to the end of the line Send a visible bell (good to send a flash to someone’s screen) Start invisible text mode Start protected mode Restore the last saved cursor position (paved by tput sc) Begin reverse video mode (bright!) End the standout mode (reverses tput smso) Ends the underline (underscore) mode Save the cursor position

tput cnorm tput cr tput cvvis tput dim tput ed tput el tput flash tput invis tput prot tput rc tput rev tput rmso tput rmul tput sc

hgrep: Highlighted grep Script
Table 15.1 tput sgr0 tput smso tput smul tput (Continued) Turn off all video modes Start the standout mode (soft reverse video we used in this chapter) Start the underline (underscore) mode Underscore one character and move to the next character

401

Table 15.1 is only an abbreviated listing of the tput command options. As you can see, we can do a lot with the text on the screen. Use your imagination, and play around with the commands.

Summary
In this chapter we introduced using reverse video to highlight text within our output. Also we showed how to do command substitution inside a sed command statement. There are many more options for the tput command to control the terminal; for example, we could have underlined the matching pattern. The nice thing about the tput command is that it will let you mix things up, too. In the next chapter we are going to look at how to keep the printers in the landscape printing. If you do not automate this function you could spend all of your time doing printer management instead of doing any real work. See you in the next chapter!

Free & Share & Open

CHAPTER

16
Print Queue Hell: Keeping the Printers Printing

If you have worked in a large systems environment for very long you already know how frustrating it can be to keep the printer farm happy. In my contracting days I worked in several shops that consistently had problems with the printers. In most cases, the print queues went down because of network timeouts and extended device waits. In this kind of environment you have two choices: keep answering the calls from the help desk or write a shell script to monitor the printer queues and reenable the queues as they drop offline. I prefer the second method. Like every other Systems Administrator, I like to be proactive in my approach to solving the little problems as well as the big ones. The shop I remember the best was a hospital. This hospital has more than 30 satellite clinics around town and only one 100MB/Sec pipe coming in to the hospital from the outside world. Most of the clinics have between three and five printers, with at least one printer active most of the day. When I came on board, the first problem I encountered was the huge volume of calls to the help desk about printer problems. What caught my eye was the fact that all of the calls came from the clinics, not from inside the hospital. I knew immediately that a shell script was in order! In this chapter we are going to look at two methods of bringing up the print queues, enabling individual queues and bringing up the whole lot. Because Unix flavors vary on handling printers and queues, we first will look at the differences between the Unix flavors.

403

Free & Share & Open

404

Chapter 16

System V versus BSD Printer Subsystems
Depending on the Unix flavor, the commands vary to control the printers and queues because some use the System V subsystem and others use BSD. With AIX you have an ever more confusing situation beginning with AIX 5L. Starting with this release, AIX now supports both the “classic” AIX printer subsystem and the System V printer service. Another problem is that some commands do not provide the full print queue name if the queue name exceeds seven characters. I have come up with some ways to get around the long queue names, and on most systems you do not have to worry about long queue names too much if you want to control all of the printers at once. In this book we are covering AIX, HP-UX, Linux, and Solaris. For no other reason that I can think of, let’s cover the printer systems in alphabetical order.

AIX Print Control Commands
AIX is the most interesting of the bunch with its new support for the System V printer service starting with AIX 5L. Although the AIX classic printer subsystem will still be supported for many years, the move seems to be going to System V for printing service.

Classic AIX Printer Subsystem
Most AIX Systems Administrators still prefer to use the classic AIX printer subsystem. This is the primary printing that I have supported for years. With the AIX printer subsystem you do not have the detailed control that the System V service offers. For example, you do not control forms and user priorities at a granular level, and you cannot manage the printers independently of the print queues easily. With this printer subsystem anyone can print on any printer, and the print queue is either UP, allowing you to print, or DOWN, disabling all printing. The shell scripts we are going to write for the classic AIX printer subsystem work at the print queue level. The two commands we are going to use are lpstat and enq -A. Both commands produce the same output, but some administrators seem to like one over the over. As I stated earlier, we need to be aware that sometimes print queues are created with queue names longer than seven characters, which is the default that can be displayed with both of these commands. I guess IBM noticed this little problem and added the -W switch to give a wide character output. Look at Listings 16.1 and 16.2 to see the different outputs.
# lpstat Queue ------hp4 hp4-ps hp4-gl yogi_hp yogi_hp Dev ----lp0 lp0 lp0 lp0 lp0 Status Job Files User PP % Blks Cp Rnk --------- --- ----------- ---------- ---- -- ----- --- --READY READY READY DOWN DOWN

Listing 16.1 Output using lpstat or enq -A.

Print Queue Hell: Keeping the Printers Printing

405

# lpstat -W Queue ---------------hp4 hp4-ps hp4-gl yogi_hp4_1 yogi_hp4_1ps

Dev ----lp0 lp0 lp0 lp0 lp0

Status Job Files User PP % Blks Cp Rnk -------- ---- -------- ------ --- -- ---- --- --READY READY READY DOWN DOWN

Listing 16.2 Output using lpstat -W or enq -AW.

As you can see in Listing 16.1, the long queue names are cut off at the seventh character when using the lpstat or enq -A commands. By adding the -W switch to these commands we see the entire long queue name. This is important because you cannot control a print queue if you do not have the exact, and full, queue name. There are two methods to script using either lpstat -W or enq -AW. One method is to loop through each queue that is reported DOWN; the other is to use one long compound command. We are first going to look at the looping method. A little for loop can be used to extract out the queue names of the printers in a DOWN state. The list used for the for loop comes from either of the following command statements: lpstat -W | tail +3 | grep DOWN | awk ‘{print $1}’

or enq -AW | tail +3 | grep DOWN | awk ‘{print $1}’

Both of the previous statements produce the same output. Notice that tail +3 is the second command in pipe, just after the lpstat and enq commands. We use tail +3 in this statement to remove the two lines of header information. This method is much cleaner than trying to grep out some unique character in both of the header lines. Notice that the number of lines, specified by +3, is one larger than the actual number of lines that we want to remove. Using the tail command this way, we are telling tail to start listing at the third line, so two lines are removed at the top of the output. The third command in the pipe is where we grep for DOWN, looking for disabled printers, as shown in Listing 16.2. The output from this stage of the command is only the lines of the enq and lpstat output that contains the word DOWN. Using these lines as input for the next command in the pipe, we are ready to extract the actual queue name(s) of the disabled printers, as shown in the output here. yogi_hp4_1 yogi_hp4_1ps lp0 lp0 DOWN DOWN

Free & Share & Open

406

Chapter 16
The awk command, as we use it, is used to extract the field that we want to work with, which is the first field, the queue name. Using the previous output as input to our awk statement we extract out the first field using the following syntax: command | awk ‘{print $1}’

You can extract any valid field using awk as well as different fields at the same time. For example, if we want to extract fields 1 and 3, specified by $1 and $3, the following awk statement will take care of the task. command | awk ‘{print $1, $3}’

Notice that I added a comma between $1 and $3. If the comma is omitted, then there will not be a space between the two strings. Instead the output will be two strings appended together without a space. For our for loop we can first send the lpstat and enq command output to a file and process the file in a loop, or we can use command substitution to add the statement directly into the for loop to create the list of objects to loop through. Let’s look at our for loop structure. for Q in $( enq -AW | tail +3 | grep DOWN | awk ‘{print $1}’ ) do # Do something here. done

Using this loop command statement, the for loop will loop through yogi_hp4_1 and yogi_hp4_1ps print queue names, which is equivalent to the following for loop structure: for Q in yogi_hp4_1 yogi_hp4_1ps do # Do something here. done

Because we never know which queues may be down, we need to parse through the output of the actual queue names of the printers in a disabled state. The shell script in its entirety is shown in Listing 16.3.
#!/bin/ksh # # SCRIPT: enable_AIX_classic.ksh # # AUTHOR: Randy Michael # DATE: 03/14/2002 # REV: 1.1.P

Listing 16.3 For loop to enable “classic” AIX print queues.

Print Queue Hell: Keeping the Printers Printing

407

# # # # # # # # # #

PLATFORM: AIX Only PURPOSE: This script is used to enable print queues on AIX systems. REV LIST: set -x # Uncomment to debug this script set -n # Uncomment to check syntax without any execution

for Q in $( enq -AW | tail +3 | grep DOWN | awk ‘{print $1}’) do enable $Q (( $? == 0 )) || echo “\n$Q print queue FAILED to enable.\n” done

Listing 16.3 For loop to enable “classic” AIX print queues. (continued)

Inside the for loop we attempt to enable each print queue individually. If the return code of the enable command is not zero we echo an error message indicating that the queue could not be enabled. Notice the highlighted lines in Listing 16.3. We use the mathematical test, specified by the double parentheses, (( math test )). Using this math test you normally do not add a dollar sign, $, in front of a numeric variable. When the variable is produced by the system, such as $?, the dollar sign is required. Testing for equality also requires using the double equal signs, ==, because the single equal sign, =, is meant as an assignment, not a test. After the test to check for a zero return code, we use a logical OR, specified by the double pipes, ||. This logical OR will execute the next command only if the return code of the enable $Q command is nonzero, which means that the command failed. There is also a logical AND that is used by placing double ampersands, &&, in a command statement. A logical AND does just the opposite; it would execute the succeeding command if the test is true, instead of false. Both the logical OR and logical AND are used as replacements for if..then..else.. statements. We can also accomplish this task by using a single compound command statement. Just as we used command substitution in the for loop, we can use command substitution to produce command parameters. For example, we can use our for loop command to create command parameters to the enable command. To see this more clearly, look at the following two commands. enable $(enq -AW | tail +3 | grep DOWN | awk ‘{print $1}’) 2>/dev/null

or enable $(lpstat -W | tail +3 | grep DOWN | awk ‘{print $1}’) 2>/dev/null

Free & Share & Open

408

Chapter 16
Both of the previous compound command statements produce the same result, enabling all of the print queues on the system. The only problem with using this technique is that if you execute this command and all of the printers are already enabled, then you will get the following output from standard error: usage: enable PrinterName ... Enables or activates printers.

As you can see, I sent this output to the bit bucket by adding 2>/dev/null to the end of the statement, but the return code is still nonzero if all of the printers are already enabled. This should not be a problem unless you want to create some notification that a printer failed to enable. In our for loop in Listing 16.3 we used the return code from the enable command to produce notification. I will leave the technique that you use up to you. If you do not want to see any output, then you could add the single compound statement as a cron table entry or use the for loop technique in a shell script to redirect the failure notification to a log file. If you use a log file you may want to add a date stamp.

System V Printing on AIX
Beginning with AIX 5L, IBM supports System V printing. I find that Solaris has the closest command usage and output. With only a few differences between AIX and Solaris System V printing in the output produced, you could use the shell scripts interchangeably. Because people tend to read only the parts of a technical book that they need to, I will devote this entire section to AIX System V printing. To switch your AIX system from the “classic” AIX printer subsystem to System V printing, refer to your AIX reference manual. This section expects that you are already running System V printing. Like Solaris, AIX uses the System V lpc (line printer control) command to control the printers and print queues. The nice thing about this print service is that you can control the queues and the printers independently. The main commands that we are interested in for AIX queuing and printing include the following options and parameters to the lpc command, as shown in Table 16.1.

Table 16.1

AIX lpc Command Options COMMAND RESULT Disables queuing Disables printing Disables printing and queuing Enables queuing Enables printing Enables printing and queuing

LPC COMMAND disable (printer[@host] | all) stop (printer[@host] | all) down (printer[@host] | all) enable (printer[@host] | all) start (printer[@host] | all) up (printer[@host] | all)

Print Queue Hell: Keeping the Printers Printing
As you can see in Table 16.1, the granularity of printer control is excellent, which gives us several options when creating shell scripts. To control all of the printing and queuing at one time you really do not need a shell script. The following two commands can start and stop all printing and queuing on all print queues at the same time. lpc down all lpc up all # Disable all printing and queuing # Enable all printing and queuing

409

To keep all of the printers printing and queuing you only need the lpc up all command entered into a cron table. I placed an entry in my root cron table to execute this lpc command every 10 minutes, as shown here:
5,15,25,35,45,55 * * * * /usr/sbin/lpc up all >/dev/null 2>&1

This cron table entry enables all printing and queuing on all printers on the 5s, 24 hours a day, 7 days a week. With AIX System V printing, the data we are interested in is separated on three lines of output when we use the lpc status all command to monitor the printer service. The same command executed on AIX , Linux, and Solaris is shown here. AIX SYSTEM V OUTPUT
# lpc status all hp4V: queueing is enabled printing is disabled 5 entries in spool area

LINUX SYSTEM V OUTPUT
# lpc status Printer Status/(Debug) hp4@localhost

Printing Spooling Jobs enabled disabled 0

Server Subserver Redirect none none

SOLARIS SYSTEM V OUTPUT
# lpc status all bambam_hp4: queueing is enabled printing is enabled no entries

Of these three outputs Linux is the one that differs. With the data we are interested in for AIX residing on three separate lines for each print queue, we need a different

Free & Share & Open

410

Chapter 16 strategy to get the exact data the we want. First notice that at the beginning of each stanza a queue name has a colon, :, appended to the name of the queue. Because this character occurs only in the queue name, we can use the colon character as a tag for a grep statement. Following the queue name entry, the next two lines contain the data that we are interested in pertaining to the status of the queuing and printing. Because we have some unique tag for each entry, it is easy to extract the lines of data that we are interested in by using an extended grep, or egrep, statement, as shown here: lpc status all | egrep ‘:|printing|queueing’ | while read LINE

The egrep command works the same way as the grep command except that you can specify multiple patterns to match. Each pattern is separated by a pipe without any spaces! If you add spaces on either side of the search pattern, the egrep statement will fail to make a match. The entire list of patterns is then enclosed within single forward tic marks, ’pattern1|pattern2|pattern3’. The output produced has the queue name on the first line, the printing status on the second line, and the queuing status on the third line. The last part of the previous command is where the output is piped to a while loop. On each read the entire line of data is loaded into the variable LINE. Inside of the while loop we use the following case statement to assign the data to the appropriate variable. case $LINE in *:) Q=$(echo $LINE | cut -d ‘:’ -f1) ;; printing*) PSTATUS=$(echo $LINE | awk ‘{print $3}’) ;; queueing*) QSTATUS=$(echo $LINE | awk ‘{print $3}’) ;; esac

Notice that if $LINE begins with *: then we load the Q variable. If $LINE begins with printing* we load the PSTATUS variable with the third field, which should be either enabled or disabled. We do the same thing in loading the QSTATUS variable with the third field of the value that the $LINE variable points to. The trick in this script is how to load and process three lines of data and then load and process three more lines of data, and so on. The most intuitive approach is to have a loop counter. Each time the loop counter reaches three we process the data and reset the loop counter back to zero. Take a look at the entire script in Listing 16.4 to see how this loop count works. Pay close attention to the bold type.

Print Queue Hell: Keeping the Printers Printing

411

#!/bin/ksh # # SCRIPT: print_UP_SYSV_AIX.ksh # # AUTHOR: Randy Michael # DATE: 03/14/2002 # REV: 1.1.P # # PLATFORM: AIX System V Printing # # PURPOSE: This script is used to enable printing and queuing separately # on each print queue on AIX and Solaris systems. # # REV LIST: # # set -x # Uncomment to debug this script # set -n # Uncomment to check syntax without any execution # ################################################# LOOP=0 # Loop Counter - To grab three lines at a time

lpc status all | egrep ‘:|printing|queueing’ | while read LINE do # Load three unique lines at a time case $LINE in *:) Q=$(echo $LINE | cut -d ‘:’ -f1) ;; printing*) PSTATUS=$(echo $LINE | awk ‘{print $3}’) ;; queueing*) QSTATUS=$(echo $LINE | awk ‘{print $3}’) ;; esac # Increment the LOOP counter (( LOOP = LOOP + 1 )) if ((LOOP == 3)) # Do we have all three lines of data? then # Check printing status case $PSTATUS in

Listing 16.4 print_UP_AIX.ksh shell script listing. (continues)

Free & Share & Open

412

Chapter 16

disabled) lpc start $Q >/dev/null (($? == 0)) && echo “\n$Q printing re-started\n” ;; enabled|*) : # No-Op - Do Nothing ;; esac # Check queuing status case $QSTATUS in disabled) lpc enable $Q >/dev/null (($? == 0)) && echo “\n$Q queueing re-enabled\n” ;; enabled|*) : # No-Op - Do Nothing ;; esac LOOP=0 # Reset the loop counter to zero fi done

Listing 16.4 print_UP_AIX.ksh shell script listing. (continued)

Notice that we grab three lines at a time. The reason that I say that we are grabbing three lines at a time is because I use the case statement to specify unique tags for each line of data. I know that the queue name will have a colon, :, as a suffix. I know that the printing status line will begin with printing*, and I know that the queuing line will begin with queueing*. We load only one variable on each loop iteration. So, to get three pieces of data (queue name, printing status, and queuing status), we need to go through the while loop three times for each printer queue. Once we pass the initial case statement, we increment the LOOP counter by one. If the $LOOP variable is equal to 3 then we have all of the data that we need to process a single printer queue. After processing the data for this printer queue, we reset the LOOP variable to zero, 0, and start gathering data for the next printer queue. Sounds simple enough? This same technique works for any fixed set of lines of data in command output or in a file. The only changes that are needed to use this method include creating unique tags for the data you are interested in and setting the $LOOP equality statement to reflect the number of lines in each set of data.

More System V Printer Commands
We have been looking at only the lpc command thus far. We also need to look at two command parameters to the lpstat command in this section. The -a parameter lists the status of queuing, and the -p command parameter lists the status of printing. The nice thing about these two command options is that the output for each queue is on a single line, which makes the data easier to parse through. See Table 16.2.

Print Queue Hell: Keeping the Printers Printing
Table 16.2 System V lpstat Command Options DESCRIPTION Show status of queuing on all printers Show status of printing on all printers

413

COMMAND lpstat -a lpstat -p

Other than having to query the printer service twice, having to use separate commands for monitoring printing and queuing is not so bad. The separation is built in because the -a and -p command parameters are mutually exclusive, which means that you cannot use -a and -p at the same time. Output from each command option is shown here:
# lpstat -a hp4 accepting requests since May 07 07:02 2002 yogi_hp4_1ps accepting requests since May 07 07:02 2002 long_queue not accepting requests since Tue May 7 07:02:23 EDT 2002 s_q_nam not accepting requests since Tue May # lpstat -p printer long_queue disabled since Tue May 7 07:02:01 EDT 2002. available. stopped by user printer s_q_nam disabled since Tue May 7 07:02:01 EDT 2002. available. stopped by user printer hp4 unknown state. enabled since May 07 07:30 2002. available. printer yogi_hp4_1ps unknown state. enabled since May 07 07:30 2002. available. 7 07:02:23 EDT 2002 -

Listing 16.5 lpstat -a and lpstat -p command output.

Notice in Listing 16.5 that the output from each command option has a unique set of status information for each printer on each line of output. We want to use the uniqueness of the status information as tags in a grep statement. The terms make sense, too. A queue is either accepting new requests or is not accepting new requests, and a printer is either enabled for printing or is disabled from printing. Because we are interested only in the disabled and not-accepting states, we can create a simple script or a one-liner. We need to know two things to enable printing and to bring up a print queue to accept new requests, the printer/queue name and the state of the queue or printer. The first step is to grep out the lines of output that contain our tag. The second step is to

Free & Share & Open

414

Chapter 16 extract the printer/queue name from each line of output. Let’s first look at using a while loop to bring everything up, as shown in the Listing 16.6. lpstat -a | grep ‘not accepting’ | while read LINE do Q=$(echo $LINE | awk ‘{print $1}’) lpc enable $Q done lpstat -p | grep disabled | while LINE do P=$(echo $LINE | awk ‘{print $2}’) lpc start $P done

Listing 16.6 Scripting the lpstat command using -a and -p.

Notice in Listing 16.6 that we have to work on the print queues and printers separately, by using two separate loops. In the first while loop all of the queuing is started. In the second loop we enable printing for each of the printers. The down side to this method occurs when you have hundreds of printers and scanning through all of the printers twice can take quite a while. Of course, if you have hundreds of printers you should use lpc up all to bring everything up at once. As I said before, we can also make a one-liner out of the two loops in Listing 16.6. We can combine the grep and awk commands on the same line and use command substitution to execute the lpc command. The following two commands replace the two while loops. lpc enable $(lpstat -a | grep ‘not accepting’ | awk ‘{print $1}’) lpc start $( lpstat -p | grep disabled | awk ‘{print $2}’)

The first command enables queuing, and the second command starts printing. The command substitution, specified by the $(command) notation, executes the appropriate lpstat command, then greps on the tag and extracts the printer/queue name out. The resulting output is used as the parameter to the lpc commands.

HP-UX Print Control Commands
Of the Unix operating systems, HP-UX has a unique lpstat command output. We do not have to do anything special to see the full print queue names, and if a queuing is disabled or printing is stopped, we get a Warning: message. With a warning message for each printer on a single line we can use grep and awk to find the printer/queue name and the status in a case statement. Let’s first look at the lpstat output when both printing and queuing is up, as shown here:

Print Queue Hell: Keeping the Printers Printing
# lpstat printer queue for hp4_yogi_1

415

printer queue for yogi_hp4_1ps

If print requests were queued up they would be listed below the queue name. Now let’s disable printing on the hp4_yogi_1 print queue.
# disable hp4_yogi_1 printer “hp4_yogi_1” now disabled

Now look at the output of the lpstat command:
# lpstat printer queue for hp4_yogi_1 dino: Warning: hp4_yogi_1 is down

printer queue for yogi_hp4_1ps

The warning message tells us that the printer is down; however, notice that the queue status is not listed here. Now let’s bring down the hp4_yogi_1 print queue and see what this does.
# reject hp4_yogi_1 destination “hp4_yogi_1” will no longer accept requests

To see only queuing status we use the lpstat -a command, as shown here:
# lpstat -a hp4_yogi_1 not accepting requests since Oct 1 05:45 reason unknown yogi_hp4_1ps accepting requests since Sep 26 04:23

Because hp4_yogi_1 now has printing disabled and queuing stopped, I would expect that we should see some queue status output in the lpstat command output for the first time.
# lpstat printer queue for hp4_yogi_1

Free & Share & Open

416

Chapter 16 dino: Warning: hp4_yogi_1 queue is turned off dino: Warning: hp4_yogi_1 is down

printer queue for yogi_hp4_1ps

Just what we expected. From this little exercise we have determined that queuing is reported only when the queuing is stopped on the queue using the lpstat command alone. For our scripting effort let’s stick to the lpstat output. We want to use the word Warning as a tag for our grep statement. Then we can further grep this extracted line to check printing and queuing status. If the string ‘queue is turned off’ is present we know that queuing is turned off, and if the string ‘is down’ appears on the line we know that printing is disabled. The only thing left to extract is the printer/queue name, which is always located in the third field. To script this we can use the code in Listing 16.7. Pay attention to the bold type, and we will cover the script at the end.
#!/bin/ksh # # SCRIPT: print_UP_HP-UX.ksh # # AUTHOR: Randy Michael # DATE: 03/14/2002 # REV: 1.1.P # # PLATFORM: HP-UX Only # # PURPOSE: This script is used to enable printing and queuing separately # on each print queue on an HP-UX system. # # REV LIST: # # set -x # Uncomment to debug this script # set -n # Uncomment to check syntax without any execution lpstat | grep Warning: | while read LINE do if (echo $LINE | grep ‘is down’) > /dev/null then enable $(echo $LINE | awk ‘{print $3}’) fi if (echo $LINE | grep ‘queue is turned off’) >/dev/null then accept $(echo $LINE | awk ‘{print $3}’) fi done

Listing 16.7 print_UP_HP-UX.ksh shell script listing.

Print Queue Hell: Keeping the Printers Printing
I want to point out a nice little trick in the shell script in Listing 16.7. In both of the if..then..fi statements, notice that we execute a command inside parentheses. What this technique allows us to do is execute a command in a sub-shell and use the command’s resulting return code directly in the if..then..fi structure. We really could not care less about seeing the line that we are grepping on; however, if the return code from the command is zero, then the pattern is present. In the first half of the script in Listing 16.7 we check the status of printing. If a printer is found to be disabled, then we use command substitution to produce the printer name for the enable command. Likewise, we check for the status of queuing in the second half of the script. Again, using command substitution we have the queue name to provide as a parameter to the accept command. Notice that I added the redirection to the bit bucket, specified by >/dev/null, after the command in the if statement. I add this redirection to /dev/null to suppress the output of the grep statement. That is it for HP-UX printing. HP did a good job of keeping everything pretty straightforward in the printing arena.

417

Linux Print Control Commands
Linux uses the System V lpc (line printer control) command to control the printers and print queues, as most System V Unix does. The nice thing about this print service is that you can control the queues and the printers independently. The main commands that we are interested in for Linux queuing and printing include the options to the lpc command listed in Table 16.3. As you can see in Table 16.3, the granularity of printer control is excellent, which gives up several options when creating shell scripts. To control all of the printing and queuing at one time you really do not need a shell script. The following two commands can start and stop all printing and queuing on all print queues at the same time. lpc down all lpc up all # Disable all printing and queuing # Enable all printing and queuing

Table 16.3

Linux lpc Command Options COMMAND RESULT Disables queuing Disables printing Disables printing and queuing Enables queuing Enables printing Enables printing and queuing

LPC COMMAND disable (printer[@host] | all) stop (printer[@host] | all) down (printer[@host] | all) enable (printer[@host] | all) start (printer[@host] | all) up (printer[@host] | all)

Free & Share & Open

418

Chapter 16
To keep all of the printers printing and queuing you need just the lpc up all command entered into a cron table. I placed an entry in my root cron table to execute this command every 10 minutes. My cron table entry is shown here:
5,15,25,35,45,55 * * * * /usr/sbin/lpc up all >/dev/null 2>&1

This cron table entry enables all printing and queuing on all printers on the 5s, 24 hours a day, 7 days a week. If you do want a little more control and if you keep a log of what is going on on a per queue/printer basis, then we have to do a little scripting. The script that follows searches all of the queues and reports on the individual status of printing and queuing and then enables each one independently. For this script we are going to use arrays to load the variables on each loop iteration. Array can be created and elements assigned values in two ways. The first technique is to use set -A to define the array and all of its elements. For example, if I want an array called QUEUE to contain the values for printing and queuing for a specified queue, I can set it up this way:
PQueue=yogi_hp4 Print_val=enabled Queue_val=disabled set -A QUEUE $PQueue $Print_val $Queue_val

We could have assigned the values directly in the set -A statement, but this time we used variables for the assignments. This statement defines an array named QUEUE that contains three array elements. The elements loaded into the array are the values that the variables $PQueue, $Print_val, and $Queue_val point to. For example, we assigned PQueue the value yogi_hp4, Print_val is assigned the value enabled, and Queue_val is assigned the value disabled. The result is that the first array element, 0 (zero) contains the value yogi_hp4, the second array element, 1 (one), has the value enabled, and the third array element, 2, contains the value disabled, which is what the $Queue_val variable points to. Using this technique requires that you access the array elements starting with 0, zero. To address the array elements you use the following syntax:
${QUEUE[0]} # Points to value assigned to the first array element, yogi_hp4 ${QUEUE[1]} # Points to value assigned to the second array element, enabled ${QUEUE[2]} # Points to the value assigned to the third array element, disabled

To address all of the array’s elements at the same time use the following syntax:
# print “${QUEUE[*]}” ----OR----

Print Queue Hell: Keeping the Printers Printing
# print “${QUEUE[@]}” yogi_hp4 enabled disabled

419

Now, before I lose you, let’s take a look at a more intuitive way of working with arrays and array elements. Instead of using the set -A command to define and load an array, we can define an array and load its elements at the same time using the following syntax:
QUEUE[1]=yogi_hp4 QUEUE[2]=enabled QUEUE[3]=disabled

Notice that the first array element is now referenced by 1, one. These commands create an array named QUEUE and load the first three array elements, referenced by 1, 2, and 3, into array QUEUE. Now you can use the array directly in a command statement by pointing to the array element that you want to use. For example, if I want to print the printing status of the yogi_hp4 print queue, I use the following syntax: echo “\nPrinter ${QUEUE[1]} has print status ${QUEUE[2]}\n”

The previous command produces the following output:
Printer yogi_hp4 has print status enabled

Now that we have seen the basics of working with arrays, let’s look at a shell script to handle keeping the printing and queuing enabled on all of the printers individually. The first step is to load an array in a while loop. This is a little different from what we did before with arrays. In this case I want to use the lpc status all command to find printers that have either printing or queuing disabled. The output of the lpc status all command is shown below.
# lpc status all Printer Status/(Debug) hp4@localhost

Printing Spooling Jobs enabled disabled 0

Server Subserver Redirect none none

This is an easy output to deal with because all of the data for each queue is on a single line. The output that we are interested in is the printer name, the printing status, and the spooling status—the first three fields on the second line. We are not interested in the first line at all so we can get rid of it with a pipe to the tail command. When we add to our command we get the following output:
# lpc status all | tail +2 yogi_hp4@localhost enabled

disabled

0

none

none

I currently have only one printer defined on this system, so the output is the status of a single printer. Now we want to load the first three fields into an array using a

Free & Share & Open

420

Chapter 16 while loop. Look at the next command line to see how we are directly loading an array called pqstat with array elements of the first three fields on each line. lpc status all | tail +2 | while read pqstat[1] pqstat[2] pqstat[3] junk

Because I want just the first three fields in the output, notice that the fourth variable in the read part of the while statement is junk. The junk variable is a catch-all variable to capture any remaining strings on the line of output in a single variable. It is a requirement that you take care of this remaining text because if you neglect adding a variable to catch any remaining characters on the line, you will read the characters in as strings on the next loop iteration! This type of error produces some strange output that is hard to find and troubleshoot. Notice that in the output of the lpc status all command the printer has queuing disabled, which is the third field. The easiest way to handle the two status fields is to use two case statements, with each tagging on a separate field. Look at the full script code in Listing 16.8, and we will cover the technique at the end.
#!/bin/ksh # # SCRIPT: print_UP_Linux.ksh # # AUTHOR: Randy Michael # DATE: 03/14/2002 # REV: 1.1.P # # PLATFORM: Linux Only # # PURPOSE: This script is used to enable printing and queuing separately # on each print queue on a Linux system. Logging can be # enabled. # # REV LIST: # # set -x # Uncomment to debug this script # set -n # Uncomment to check syntax without any execution # ################################################# # Initial Variables Here ################################################# LOGILE=/usr/local/log/PQlog.log [ -f $LOGFILE ] || echo /dev/null > $LOGFILE ################################################# lpc status | tail +2 | while read pqstat[1] pqstat[2] pqstat[3] junk

Listing 16.8 print_UP_Linux.ksh shell script listing.

Print Queue Hell: Keeping the Printers Printing

421

do # First check the status of printing for each printer case ${pqstat[2]} in disabled) # Printing is disabled - print status and restart printing echo “${pqstat[1]} Printing is ${pqstat[2]}” \ | tee -a$LOGFILE lpc start ${pqstat[1]} | tee -a $LOGFILE (($? == 0)) && echo “${pqstat[1]} Printing Restarted” \ | tee -a $LOGFILE ;; enabled|*) : # No-Op - Do Nothing ;; esac # Next check the status of queueing for each printer case ${pqstat[3]} in disabled) echo “${pqstat[1]} Queueing is ${pqstat[3]}” \ | tee -a $LOGFILE lpc enable ${pqstat[1]} | tee -a $LOGFILE (($? == 0)) && echo “${pqstat[1]} Printing Restarted” \ | tee -a $LOGFILE ;; enabled|*) : # No-Op - Do Nothing ;; esac done

Listing 16.8 print_UP_Linux.ksh shell script listing. (continued)

We start off this script in Listing 16.8 by defining the $LOGFILE. Notice that the following command, after the log file definition, checks to see if the log file exists. If the $LOGFILE does not exist, then the result of the test is a nonzero return code. We use a logical OR, specified by the double pipes, ||, to execute the succeeding command to create a zero length $LOGFILE because it does not exist if the return code of the test is nonzero. Next, we start our while loop to load the pqstat array on each loop iteration, which in our case is a single loop iteration for a single printer. This means that we load a one-dimensional array with new data on each loop iteration (one-dimensional arrays are all that the Korn shell can use). Again, notice the junk variable that is added as the last variable in the while loop statement. This extra variable is required to catch the remaining text in a single variable. With the array loaded we proceed with two case statements to test for the status of printing and queuing on each print queue. Notice that we use the array element directly in the case statement, as shown here: case ${pqstat[2]} in

Free & Share & Open

422

Chapter 16
We use the same technique with the print queuing array element in a separate case statement. We have only two possible results for the array elements, enabled and disabled. The only result we are concerned about is any disabled value. If we receive any disabled values we attempt to reenable the printing or queuing on the printer. Notice that the second option in both case statements includes enabled and anything else, specified by the wildcard, *, as shown here: enabled|*) We could have just used the wildcard to cover everything, but it is clearer to the reader of the script to see actual expected results in a case statement than just a catchall asterisk. When a reenabling task is completed successfully, notice the use of the logical AND to test the return code and give notification on a zero return code value, as shown here:
(($? == 0)) && echo “${pqstat[1]} Printing Restarted”

The second part of the command will execute only if the test for a zero return code is true. Otherwise, the system will report an error, so there is no need for us to add any failure notification. To see everything that is happening on the screen and to log everything at the same time we use the tee -a command. This command works with a pipe and prints all of the output to the screen; at the same time it sends the exact same output to the file specified after tee -a. An example is shown here. lpc start ${pqstat[1]} | tee -a $LOGFILE

The previous command attempts to restart printing on the print queue specified by the array element pqstat[1] and sends any resulting output to the screen and to the $LOGFILE simultaneously.

Controlling Queuing and Printing Individually
Depending on the situation, you may not always want to enable printing and queuing at the same time. We can break up the shell script in Listing 16.8 and pull out the individual case statements to start either printing or queuing. Because printing is controlled by array element 2 we can extract the first case statement to create a new shell script. Let’s call this shell script printing_only_UP_Linux.ksh. You can see the modifications in Listing 16.9.
#!/bin/ksh # # SCRIPT: printing_only_UP_Linux.ksh #

Listing 16.9 printing_only_UP_Linux.ksh shell script listing.

Print Queue Hell: Keeping the Printers Printing

423

# AUTHOR: Randy Michael # DATE: 03/14/2002 # REV: 1.1.P # # PLATFORM: Linux Only # # PURPOSE: This script is used to enable printing on each printer # on a Linux system. Logging is enabled. # # REV LIST: # # set -x # Uncomment to debug this script # set -n # Uncomment to check syntax without any execution # ################################################# # Initial Variables Here ################################################# LOGILE=/usr/local/log/PQlog.log [ -f $LOGFILE ] || echo /dev/null > $LOGFILE ################################################# lpc status | tail +2 | while read pqstat[1] pqstat[2] pqstat[3] junk do # Check the status of printing for each printer case ${pqstat[2]} in disabled) # Printing is disabled - print status and restart printing echo “${pqstat[1]} Printing is ${pqstat[2]}” \ | tee -a$LOGFILE lpc start ${pqstat[1]} | tee -a $LOGFILE (($? == 0)) && echo “${pqstat[1]} Printing Restarted” \ | tee -a $LOGFILE ;; enabled|*) : # No-Op - Do Nothing ;; esac done

Listing 16.9 printing_only_UP_Linux.ksh shell script listing. (continued)

Notice that the only thing that was changed is that the second case statement structure was removed from the script and the name was changed. We can do the same thing to create a shell script that only enables queuing, as shown in Listing 16.10.

Free & Share & Open

424

Chapter 16

#!/bin/ksh # # SCRIPT: queuing_only_UP_Linux.ksh # # AUTHOR: Randy Michael # DATE: 03/14/2002 # REV: 1.1.P # # PLATFORM: Linux Only # # PURPOSE: This script is used to enable printing and queuing separately # on each print queue on a Linux system. Logging can be # enabled. # # REV LIST: # # set -x # Uncomment to debug this script # set -n # Uncomment to check syntax without any execution # ################################################# # Initial Variables Here ################################################# LOGILE=/usr/local/log/PQlog.log [ -f $LOGFILE ] || echo /dev/null > $LOGFILE ################################################# lpc status | tail +2 | while read pqstat[1] pqstat[2] pqstat[3] junk do # check the status of queueing for each printer case ${pqstat[3]} in disabled) echo “${pqstat[1]} Queueing is ${pqstat[3]}” \ | tee -a $LOGFILE lpc enable ${pqstat[1]} | tee -a $LOGFILE (($? == 0)) && echo “${pqstat[1]} Printing Restarted” \ | tee -a $LOGFILE ;; enabled|*) : # No-Op - Do Nothing ;; esac done

Listing 16.10 queuing_only_UP_Linux.ksh shell script listing.

Notice that the only thing that was changed this time is the first case statement structure was removed from the script and the name of the shell script was changed.

Print Queue Hell: Keeping the Printers Printing
You could also modify the shell script in Listing 16.8 to add a command-line parameter to let you control queuing and printing individually from the same shell script. I am going to leave this as an exercise for you to complete. As a hint for this exercise: Expect only zero or one command-line parameters. If $# is equal to zero, then enable both queuing and printing. If there is one parameter and the value of $1 is “all”, then enable both printing and queuing. If the $1 parameter is equal to “printing”, then enable only printing. If $1 is equal to “queuing”, then enable only queuing. You need to add a usage function to show how to use the shell script if the given value does not match what you are expecting. Arrays are good to use in a lot of situations where you want to address certain output fields directly and randomly. All Korn shell arrays are one-dimensional arrays, but using the array in a loop gives the appearance of a two-dimensional array.

425

Solaris Print Control Commands
Solaris uses the System V lpc (line printer control) command to control the printers and print queues, as most System V Unix does. The nice thing about this print service is that you can control the queues and the printers independently. The main commands that we are interested in for Solaris queuing and printing include the following options and parameters to the lpc command, as shown in Table 16.4. As you can see in Table 16.4, the granularity of printer control is excellent, which gives several options when creating shell scripts. To control all of the printing and queuing at one time you really do not need a shell script. The following two commands can start and stop all printing and queuing on all print queues at the same time. lpc down all lpc up all # Disable all printing and queuing # Enable all printing and queuing

To keep all of the printers printing and queuing you need only the lpc up all command entered into a cron table. I placed an entry in my root cron table to execute this command every 10 minutes. My cron table entry is shown here:
5,15,25,35,45,55 * * * * /usr/sbin/lpc up all >/dev/null 2>&1

Table 16.4

Solaris lpc Command Options COMMAND RESULT Disables queuing Disables printing Disables printing and queuing Enables queuing Enables printing Enables printing and queuing

LPC COMMAND disable (printer[@host] | all) stop (printer[@host] | all) down (printer[@host] | all) enable (printer[@host] | all) start (printer[@host] | all) up (printer[@host] | all)

Free & Share & Open

426

Chapter 16
This cron table entry enables all printing and queuing on all printers on the 5s, 24 hours a day, 7 days a week. We have a nice situation here because we can use the same shell script that we used for the AIX System V printing on Solaris. Unlike Linux, where all of the data that we want is on a single line of output, with Solaris and AIX System V printing, the data we are interested in is separated on three lines of output. You can see the difference in the output here. AIX SYSTEM V OUTPUT
# lpc status all hp4V: queueing is enabled printing is disabled 5 entries in spool area

LINUX SYSTEM V OUTPUT
# lpc status Printer Status/(Debug) hp4@localhost

Printing Spooling Jobs enabled disabled 0

Server Subserver Redirect none none

SOLARIS SYSTEM V OUTPUT
# lpc status all bambam_hp4: queueing is enabled printing is enabled no entries

Of these three outputs, Linux is the one that differs. With the data we are interested in for Solaris residing on three separate lines for each print queue, we need a different strategy to get the exact data the we want. First notice that the beginning of the stanza for the queue name there is a colon, :, appended to the name of the queue. Because this character occurs only in the queue name, we can use the colon character as a tag for a grep statement. Following the queue name entry the next two lines contain the data pertaining to the status of the queuing and printing. Because we have some unique tag for each entry, it is easy to extract the lines of data that we are interested in by using an extended grep, or egrep, statement, as shown here: lpc status all | egrep ‘:|printing|queueing’ | while read LINE

The egrep command works the same way as the grep command except that you can specify multiple patterns to match. Each pattern is separated by a pipe without any

Print Queue Hell: Keeping the Printers Printing spaces! If you add spaces on either side of the search pattern the egrep statement will fail to make a match. The entire list of patterns is then enclosed within single forward tic marks, ’pattern1|pattern2|pattern3’. The output produced has the queue name on the first line, the printing status on the second line, and the queuing status on the third line. The last part of the previous command is where the output is piped to a while loop. On each read, the entire line of data is loaded into the variable LINE. Inside of the while loop we use the following case statement to assign the data to the appropriate variable. case $LINE in *:) Q=$(echo $LINE | cut -d ‘:’ -f1) ;; printing*) PSTATUS=$(echo $LINE | awk ‘{print $3}’) ;; queueing*) QSTATUS=$(echo $LINE | awk ‘{print $3}’) ;; esac

427

Notice that if $LINE begins with *: then we load the Q variable. If $LINE begins with printing* we load the PSTATUS variable with the third field, which should be either enabled or disabled. We do the same thing in loading the QSTATUS variable with the third field of the value that the $LINE variable points to. The trick in this script is how to load and process three lines of data and then load and process three more lines of data, and so on. The most intuitive approach is to have a loop counter. Each time the loop counter reaches three we process the data and reset the loop counter back to zero. Take a look at the entire script in Listing 16.11 to see how this loop count works. Pay close attention to the bold type.
#!/bin/ksh # # SCRIPT: print_UP_Solaris.ksh # # AUTHOR: Randy Michael # DATE: 03/14/2002 # REV: 1.1.P # # PLATFORM: Solaris Only # # PURPOSE: This script is used to enable printing and queuing separately # on each print queue on Solaris systems. # # # REV LIST:

Listing 16.11 print_UP_SUN.ksh shell script listing. (continues)

Free & Share & Open

428

Chapter 16

# # set -x # Uncomment to debug this script # set -n # Uncomment to check syntax without any execution # ################################################# LOOP=0 # Loop Counter - To grab three lines at a time

lpc status all | egrep ‘:|printing|queueing’ | while read LINE do # Load three unique lines at a time case $LINE in *:) Q=$(echo $LINE | cut -d ‘:’ -f1) ;; printing*) PSTATUS=$(echo $LINE | awk ‘{print $3}’) ;; queueing*) QSTATUS=$(echo $LINE | awk ‘{print $3}’) ;; esac # Increment the LOOP counter (( LOOP = LOOP + 1 )) if ((LOOP == 3)) # Do we have all three lines of data? then # Check printing status case $PSTATUS in disabled) lpc start $Q >/dev/null (($? == 0)) && echo “\n$Q printing re-started\n” ;; enabled|*) : # No-Op - Do Nothing ;; esac # Check queuing status case $QSTATUS in disabled) lpc enable $Q >/dev/null (($? == 0)) && echo “\n$Q queueing re-enabled\n” ;; enabled|*) : # No-Op - Do Nothing ;; esac LOOP=0 # Reset the loop counter to zero fi done

Listing 16.11 print_UP_SUN.ksh shell script listing. (continued)

Print Queue Hell: Keeping the Printers Printing
Table 16.5 System V lpstat Command Options DESCRIPTION Show status of queuing on all printers Show status of printing on all printers

429

COMMAND lpstat -a lpstat -p

Within this while loop we are grabbing three lines of data at a time to process. I say that we are grabbing three lines at a time in Listing 16.11 because I use the case statement to specify unique tags for each line of data. I know that the queue name will have a colon, :, as a suffix. I know that the printing status line will begin with printing*, and I know that the queuing line will begin with queueing*. We load only one variable on each loop iteration, though. To get three pieces of data (queue name, printing status, and queuing status), we need to go through the while loop three times for each printer queue. Once we pass the initial case statement we increment the LOOP counter by one. If the $LOOP variable is equal to 3, then we have all data that we need to process a single printer queue. After processing the data for this printer queue we reset the LOOP variable to zero, 0, and start gathering data for the next printer queue. Sounds simple enough? This same technique works for any fixed set of lines of data in command output or in a file. The only changes that are needed to use this method include creating unique tags for the data you are interested in and setting the $LOOP equality statement to reflect the number of lines that are in each set of data.

More System V Printer Commands
We have been looking only at the lpc command thus far. We also need to look at two command parameters to the lpstat command in this section. The -a parameter lists the status of queuing, and the -p command parameter lists the status of printing. The nice thing about these two command options is that the output for each queue is on a single line, which makes the data easier to parse through. The lpstat command options are shown in Table 16.5. Other than having to query the printer subsystem twice, having to use separate commands for monitoring printing and queuing is not so bad. The separation is built in because the -a and -p command parameters are mutually exclusive, which means that you cannot use -a and -p at the same time. Output from each command option is shown here:
# lpstat -a hp4 accepting requests since May 07 07:02 2002 yogi_hp4_1ps accepting requests since May 07 07:02 2002

Listing 16.12 lpstat -a and lpstat -p command output. (continues)

Free & Share & Open

430

Chapter 16

long_queue not accepting requests since Tue May s_q_nam not accepting requests since Tue May # lpstat -p

7 07:02:23 EDT 2002 -

7 07:02:23 EDT 2002 -

printer long_queue disabled since Tue May 7 07:02:01 EDT 2002. available. stopped by user printer s_q_nam disabled since Tue May 7 07:02:01 EDT 2002. available. stopped by user printer hp4 unknown state. enabled since May 07 07:30 2002. available. printer yogi_hp4_1ps unknown state. enabled since May 07 07:30 2002. available.

Listing 16.12 lpstat -a and lpstat -p command output. (continued)

Notice in Listing 16.12 that the output from each command option has a unique set of status information for each printer on each line of output. We want to use the uniqueness of the status information as tags in a grep statement. The terms make sense, too. A queue is either accepting new requests or not accepting new requests, and a printer is either enabled for printing or disabled from printing. Because we are interested in only the disabled and not accepting states, we can create a simple script or a oneliner. We need to know two things to enable printing and to bring up a print queue to accept new requests, the printer/queue name and the state of the queue or printer. The first step is to grep out the lines of output that contain our tag. The second step is to extract the printer/queue name from each line of output. Let’s first look at using a while loop to bring everything up, as shown in Listing 16.13. lpstat -a | grep ‘not accepting’ | while read LINE do Q=$(echo $LINE | awk ‘{print $1}’) lpc enable $Q done lpstat -p | grep disabled | while LINE do P=$(echo $LINE | awk ‘{print $2}’) lpc start $P done

Listing 16.13 Scripting the lpstat command using -a and -p.

Print Queue Hell: Keeping the Printers Printing
Notice in Listing 16.13 that we have to work on the print queues and printers separately, by using two separate loops. In the first while loop all of the queuing is started. In the second loop we enable printing for each of the printers. The down side to this method is where you have hundreds of printers. The time it takes to scan through all of the printers once and then rescan the printer service can be quite long. Of course, if you have hundreds of printers, you should use lpc up all to bring everything up at once. As I said before, we can also make a one-liner out of the two loops in Listing 16.13. We can combine the grep and awk commands on the same line and use command substitution to execute the lpc command. The following two commands replace the two while loops. lpc enable $(lpstat -a | grep ‘not accepting’ | awk ‘{print $1}’) lpc start $( lpstat -p | grep disabled | awk ‘{print $2}’)

431

The first command enables queuing, and the second command starts printing. The command substitution, specified by the $(command) notation, executes the appropriate lpstat command, then greps on the tag and extracts the printer/queue name. The resulting output is used as the parameter to the lpc commands.

Putting It All Together
Now we need to combine the shell scripts for each of the different Unix flavors so that one script does it all. Please do not think that taking several shell scripts, making functions out of them, and combining the new functions into a new script are difficult tasks. To make one script out of this chapter we are going to take the best of our scripts and extract the code. For each shell script we make a new function, which requires only the word function, a function name, and the code block surrounded by curly braces, function function_name { code stuff here }. Let’s take a look at the entire combined shell script in Listing 16.14 and cover the functions at the end.
#!/bin/ksh # # SCRIPT: PQ_all_in_one.ksh # # AUTHOR: Randy Michael # DATE: 03/14/2002 # REV: 1.1.P # # PLATFORM: AIX, HP-UX, Linux and Solaris

Listing 16.14 PQ_all_in_one.ksh shell script listing. (continues)

Free & Share & Open

432

Chapter 16

# # PURPOSE: This script is used to enable printing and queuing on # AIX, HP-UX, Linux and Solaris # # REV LIST: # # set -x # Uncomment to debug this script # set -n # Uncomment to check syntax without any execution # ################################################# ############### DEFINE FUNCTIONS HERE ################# function AIX_classic_printing { for Q in $( enq -AW | tail +3 | grep DOWN | awk ‘{print $1}’) do enable $Q (( $? == 0 )) || echo “\n$Q print queue FAILED to enable.\n” done } ####################################################### function AIX_SYSV_printing { LOOP=0 # Loop Counter - To grab three lines at a time lpc status all | egrep ‘:|printing|queueing’ | while read LINE do # Load three unique lines at a time case $LINE in *:) Q=$(echo $LINE | cut -d ‘:’ -f1) ;; printing*) PSTATUS=$(echo $LINE | awk ‘{print $3}’) ;; queueing*) QSTATUS=$(echo $LINE | awk ‘{print $3}’) ;; esac # Increment the LOOP counter (( LOOP = LOOP + 1 )) if ((LOOP == 3)) # Do we have all three lines of data? then # Check printing status case $PSTATUS in

Listing 16.14 PQ_all_in_one.ksh shell script listing. (continued)

Print Queue Hell: Keeping the Printers Printing

433

disabled) lpc start $Q >/dev/null (($? == 0)) && echo “\n$Q printing re-started\n” ;; enabled|*) : # No-Op - Do Nothing ;; esac # Check queuing status case $QSTATUS in disabled) lpc enable $Q >/dev/null (($? == 0)) && echo “\n$Q queueing re-enabled\n” ;; enabled|*) : # No-Op - Do Nothing ;; esac LOOP=0 # Reset the loop counter to zero fi done } ####################################################### function HP_UX_printing { lpstat | grep Warning: | while read LINE do if (echo $LINE | grep ‘is down’) > /dev/null then enable $(echo $LINE | awk ‘{print $3}’) fi if (echo $LINE | grep ‘queue is turned off’) >/dev/null then accept $(echo $LINE | awk ‘{print $3}’) fi done } ####################################################### function Linux_printing { lpc status | tail +2 | while read pqstat[1] pqstat[2] pqstat[3] junk do # First check the status of printing for each printer case ${pqstat[2]} in

Listing 16.14 PQ_all_in_one.ksh shell script listing. (continues)

Free & Share & Open

434

Chapter 16

disabled) # Printing is disabled - print status and restart printing echo “${pqstat[1]} Printing is ${pqstat[2]}” lpc start ${pqstat[1]} (($? == 0)) && echo “${pqstat[1]} Printing Restarted” ;; enabled|*) : # No-Op - Do Nothing ;; esac # Next check the status of queueing for each printer case ${pqstat[3]} in disabled) echo “${pqstat[1]} Queueing is ${pqstat[3]}” lpc enable ${pqstat[1]} (($? == 0)) && echo “${pqstat[1]} Printing Restarted” ;; enabled|*) : # No-Op - Do Nothing ;; esac done } ####################################################### function Solaris_printing { LOOP=0 # Loop Counter - To grab three lines at a time lpc status all | egrep ‘:|printing|queueing’ | while read LINE do # Load three unique lines at a time case $LINE in *:) Q=$(echo $LINE | cut -d ‘:’ -f1) ;; printing*) PSTATUS=$(echo $LINE | awk ‘{print $3}’) ;; queueing*) QSTATUS=$(echo $LINE | awk ‘{print $3}’) ;; esac # Increment the LOOP counter (( LOOP = LOOP + 1 )) if ((LOOP == 3)) # Do we have all three lines of data? then

Listing 16.14 PQ_all_in_one.ksh shell script listing. (continued)

Print Queue Hell: Keeping the Printers Printing

435

# Check printing status case $PSTATUS in disabled) lpc start $Q >/dev/null (($? == 0)) && echo “\n$Q printing re-started\n” ;; enabled|*) : # No-Op - Do Nothing ;; esac # Check queuing status case $QSTATUS in disabled) lpc enable $Q >/dev/null (($? == 0)) && echo “\n$Q queueing re-enabled\n” ;; enabled|*) : # No-Op - Do Nothing ;; esac LOOP=0 # Reset the loop counter to zero fi done } ####################################################### ############### BEGINNING OF MAIN ##################### ####################################################### # What OS are we running? # # # # To start with we need to know the Unix flavor. This case statement runs the uname command to determine the OS name. Different functions are used for each OS to restart printing and queuing.

case $(uname) in AIX) # AIX okay...Which printer subsystem? # Starting with AIX 5L we support System V printing also! # Check for an active qdaemon using the SRC lssrc command if (ps -ef | grep ‘/usr/sbin/qdaemon’ | grep -v grep) \ >/dev/null 2>&1 then # Standard AIX printer subsystem found AIX_PSS=CLASSIC elif (ps -ef | grep ‘/usr/lib/lp/lpsched’ | grep -v grep) \ >/dev/null 2>&1

Listing 16.14 PQ_all_in_one.ksh shell script listing. (continues)

Free & Share & Open

436

Chapter 16

then # AIX System V printer service is running AIX_PSS=SYSTEMV fi # Call the correct function for Classic AIX or SysV printing case $AIX_PSS in CLASSIC) # Call the classic AIX printing function AIX_classic_printing ;; SYSTEMV) # Call the AIX SysV printing function AIX_SYSV_printing ;; esac ;; # Call the HP-UX printing function HP_UX_printing ;; # Call the Linux printing function Linux_printing

HP-UX)

Linux)

;; SunOS) # Call the Solaris printing function Solaris_printing ;; *) # Anything else is unsupported. echo “\nERROR: Unsupported Operating System: $(uname)\n” echo “\n\t\t...EXITING...\n” ;;

esac

Listing 16.14 PQ_all_in_one.ksh shell script listing. (continued)

For each of the operating systems and, in the case of AIX, each printer service we took the previously created shell scripts, extracted the code, and placed it between function function_name { and the function ending character }. We now have the following functions:
AIX_classic_printing AIX_SYSV_printing HP_UX_printing Linux_printing Solaris_printing

Print Queue Hell: Keeping the Printers Printing
To execute the correct function for a specific operating system, we need to know the Unix flavor. The uname command returns the following output for each of our target operating systems:
OS AIX HP-UX Linux Solaris uname Output AIX HP-UX Linux SunOS

437

With the exception of AIX, this information is all that is needed to execute the correct function. But with AIX we have to determine which printer service is running on the server. Both types of print services have a process controlling them so we can grep for each of the processes using the ps -ef command to find the currently running printer service. When the classic AIX printer subsystem is running, there is a /usr/ sbin/qdaemon process running. When the System V printer service is running, there is a /usr/lib/lp/lpsched process running. With this information we have everything needed to make a decision on the correct function to run. We added at the end of the script all of the function execution control in the case statement that is shown in Listing 16.15. case $(uname) in AIX) # AIX okay...Which printer subsystem? # Starting with AIX 5L we support System V printing also! # Check for an active qdaemon using the SRC lssrc command if (ps -ef | grep ‘/usr/sbin/qdaemon’ | grep -v grep) \ >/dev/null 2>&1 then # Standard AIX printer subsystem found AIX_PSS=CLASSIC elif (ps -ef | grep ‘/usr/lib/lp/lpsched’ | grep -v grep) \ >/dev/null 2>&1 then # AIX System V printer service is running AIX_PSS=SYSTEMV fi # Call the correct function for Classic AIX or SysV printing case $AIX_PSS in CLASSIC) # Call the classic AIX printing function AIX_classic_printing ;;

Listing 16.15 Controlling case statement listing to pick the OS. (continues)

Free & Share & Open

438

Chapter 16

SYSTEMV)

# Call the AIX SysV printing function AIX_SYSV_printing ;;

esac ;; # Call the HP-UX printing function HP_UX_printing ;; Linux) # Call the Linux printing function Linux_printing ;; SunOS) # Call the Solaris printing function Solaris_printing

HP-UX)

*)

;; # Anything else is unsupported. echo “\nERROR: Unsupported Operating System: $(uname)\n” echo “\n\t\t...EXITING...\n” ;;

esac

Listing 16.15 Controlling case statement listing to pick the OS. (continued)

I hope by now that the code in the case statement is intuitively obvious to read and understand. If not, the first line of the case block of code is the uname command. At this point we know what the OS flavor is. For HP-UX, Linux, and Solaris we execute the target OS printing function. For AIX we make an additional test to figure out which one of the supported printing services is running. The two options are System V and the Classic AIX printer subsystem. Notice that I removed all of the logging functionality from the functions. With this type of setup, where you have the functions doing the work, you can move the logging out to the main body of the shell script. This means that you can capture all of the output data of the function to save to a log file, use the tee command to view the data while logging at the same time, or just point it to the bit bucket by redirection to /dev/null.

Other Options to Consider
As usual, we can always improve on a shell script, and these shell scripts are no exception. Some options that you may want to consider are listed next.

Print Queue Hell: Keeping the Printers Printing

439

Logging
You may want to add logging with date/time stamps. If you are having a lot of trouble keeping certain print queues up, studying the log may give you a trend that can help you find the cause of the problem. Some queues may drop in a particular location more than others. This can indicate network problems to the site. Any time you start logging do not forget to keep an eye on the log files! I often see that a script is added to a production machine, and the next thing you know, the log file has grown so large that it has filled up the filesystem. Don’t forget to prune the log files. Trimming the log files is another little shell script for you to write.

Exceptions Capability
In a lot of shops you do not want to enable every single printer and print queue. In this case you can create an exceptions file, which contains the queue/printer names that you want to exclude from enabling. You also may have special considerations if your shop uses specific forms at different times on some of the floating printers. Some shops are just print queue hell! Having the capability to keep the majority of the printers active all of the time and exclude a few is a nice thing to have.

Maintenance
During maintenance windows and other times when you want to stop all printing, you may want to comment out any cron table entries that are executing the enabling scripts. You usually find this out after the fact.

Scheduling
I keep a script running 24 × 7 to keep all of the printers available. You may want to tailor the monitoring scheduling to fit business hours (my requirement is 24 × 7). Users’ loading up on print jobs during the middle of the day is always a problem, so we try to hold big jobs for times of low activity. Low activity times are the times when you want to be at home so make sure you are keeping the printers printing during these hours, or the next morning you will have the same problem.

Summary
In this chapter we covered some unique techniques to handle the data from command output. In the Linux script we used arrays to hold the data as array elements. In other cases we read in a line at a time and used tags to grab the data we needed. We learned how to process a specific number of lines of data in groups by using a loop counter within a while loop.

Free & Share & Open

440

Chapter 16
The techniques in this chapter are varied, but the solutions are readable and can be easily maintained. Someone will follow in your footsteps and try to figure out what you did when you wrote the shell script. Do not play the “job security” game because you are you own worst enemy when it comes to documenting your shell scripts. If you comment when you write the script and make a note in the REV section when you edit it, you will have a long, happy life using your shell script. In the next chapter we are going to move into the world of FTP. The object of the next chapter is to automate file transfers between systems using FTP, or file transfer protocol.

CHAPTER

17
Automated FTP Stuff

In many shops the business relies on nightly, or even hourly, file transfers of data that is to be processed. Due to the importance of this data, the data movement must be automated. The extent of automation in the ftp world is threefold. We want the ability to move outbound files to another site, move inbound files from a remote location to your local machine, and check a remote site on a regular basis for files that are ready to download. In this chapter we are going to create some shell scripts to handle each of these scenarios. Most businesses that rely on this type of data movement also require some pre-ftp and post-ftp processing to ready the system for the files before the transfer takes place and to verify the data integrity or file permissions after the transfer. For this pre and post processing we need to build into the shell script the ability to either hard-code the pre and post processing events or point to a file that performs these tasks. Now we are up to five pieces of code that we need to create. Before we go any further let’s look at the syntax for the ftp connections.

Syntax
Normally when we ftp a file, the remote machine’s hostname is included as an argument to the ftp command. We are prompted for the password and, if it is entered correctly, we are logged into the remote machine. We then can move to the local directory

441

Free & Share & Open

442

Chapter 17 containing the file we want to upload, then to the directory that is to receive the upload from our local machine. In either case we are working with an interactive program. A typical ftp session looks like the output shown in Listing 17.1.
[root:yogi]@/# cd /scripts/download [root:yogi]@/scripts/download# ftp wilma Connected to wilma. 220 wilma FTP server (SunOS 5.8) ready. Name (wilma:root): randy 331 Password required for randy. Password: 230 User randy logged in. ftp> cd /scripts 250 CWD command successful. ftp> get auto_ftp_xfer.ksh 200 PORT command successful. 150 ASCII data connection for auto_ftp_xfer.ksh (10.10.10.1,32787) (227 bytes). 226 ASCII Transfer complete. 246 bytes received in 0.0229 seconds (10.49 Kbytes/s) local: auto_ftp_xfer.ksh remote: auto_ftp_xfer.ksh ftp> bye 221 Goodbye. [root:yogi]@/scripts/download#

Listing 17.1 Typical FTP file download.

As you can see in Listing 17.1 the ftp command requires interaction with the user to make the transfer of the file from the remote machine to the local machine. How do we automate this interactive process? If you have been studying other chapters, then you know the answer is a here document. A here document is a coding technique that allows us to place all of the required interactive command input between two labels. Let’s look at an example of coding a simple ftp transfer using this automation technique in Listing 17.2.
#!/bin/ksh # # SCRIPT: tst_ftp.ksh # AUTHOR: Randy Michael # DATE: 6/12/2002 # REV: 1.1.A # PLATOFRM: Not platform dependent

Listing 17.2 Simple here document for FTP transfer in a script.

Automated FTP Stuff

443

# # PURPOSE: This shell script is a simple demonstration of # using a here document in a shell script to automate # an FTP file transfer. # # Connect to the remote machine and begin a here document. ftp -i -v -n wilma $SEARCH_PATH” >> $OUTFILE “\nDate/Time of Search: `date`” >> $OUTFILE “\nSearch Results Sorted by File Modification Time” >> $OUTFILE

############################################

Listing 18.1 findlarge.ksh shell script listing. (continues)

Free & Share & Open

470

Chapter 18

# Search for files > $MEG_BYTES starting at the $SEARCH_PATH # find $SEARCH_PATH -type f -size +${MEG_BYTES}000000c \ -print > $HOLDFILE # How many files were found? if [ -s $HOLDFILE ] # File greater than zero bytes? then NUMBER_OF_FILES=`cat $HOLDFILE | wc -l` echo “\nNumber of Files Found: ==> $NUMBER_OF_FILES\n\n” >> $OUTFILE # Append to the end of the Output File... ls -lt `cat $HOLDFILE` >> $OUTFILE # Display the Time Sorted Output File... more $OUTFILE echo “\nThese Search Results are Stored in ==> $OUTFILE” echo “\nSearch Complete...EXITING...\n” else cat $OUTFILE # Show the header information! echo “\n\nNo Files were Found in the Search Path that” echo “are Larger than ${MEG_BYTES}Mb\n” echo “\n\t...EXITING...\n” fi rm -f $HOLDFILE # Remove the temp. file

# End of the findlarge.ksh Script

Listing 18.1 findlarge.ksh shell script listing. (continued)

Let’s review the findlarge.ksh shell script in Listing 18.1 in a little more detail. We added two functions to our script. We always need a usage function, and in case CTRL-C is pressed we added a trap_exit function. The trap_exit function is executed by the trap for exit signals 1, 2, 3, and 15 and will display EXITING ON A TRAPPED SIGNAL before exiting with a return code of 2. The usage function is executed if any of our three previously discussed data tests fail and the script exits with a return code of 1, one, indicating a script usage error. In the next block of code we query the system for the hostname, date/time stamp, and the search path (the current directory!) for the find command. All of this system data is used in the file header for the $OUTFILE. For the search path we could have just

Finding “Large” Files used a dot to specify the current directory, but this short notation would result in a relative pathname in our report. The full pathname, which begins with a forward slash (/), provides much clearer information and results in an easier-to-read file report. To get the full pathnames for our report, we use the pwd command output assigned to the SEARCH_PATH variable. We define two files for processing the data. The $HOLDFILE holds the search results of the find command’s output. The $OUTFILE contains the header data, and the search results of the find command are appended to the end of the $OUTFILE file. If the $HOLDFILE is zero-sized, then the find command did not find any files larger than $MEG_BYTES, which is the value specified in $1 on the command line. If the $HOLDFILE is not empty, we count the lines in the file with the command NUMBER_OF_ LINES=`cat $HOLDFILE | wc -l`. Notice that we used back tics for command substitution, `command`. This file count is displayed along with the report header information in our output file. The search data from the find command, stored in $HOLDFILE, consists of full pathnames of each file that has exceeded our limit. In the process of appending the $HOLDFILE data to our $OUTFILE, we do a long listing sorted by the modification time of each file. This long listing is produced using the ls -lt $(cat $HOLDFILE) command. A long listing is needed in the report so that we can see not only the modification date/time but also the file owner and group as well as the size of each file. All of the data in the $OUTFILE is displayed by using the more command so that we display the data one page at a time. The findlarge.ksh shell script is in action in Listing 18.2.
Searching for Files Larger Than 1Mb starting in: ==> /scripts Please Standby for the Search Results... Large Files Search Results: Hostname of Machine: yogi Top Level Directory of Search: ==> /scripts Date/Time of Search: Thu Nov 8 10:46:21 EST 2001

471

Search Results Sorted by File Modification Time: Number of Files Found: ==> -rwxrwxrwx 1 root sys /scripts/sling_shot621.tar 4 3490332 Oct 25 10:03

Listing 18.2 findlarge.ksh shell script in action. (continues)

Free & Share & Open

472

Chapter 18

-rwxrwxrwx 1 root sys 1280000 Aug 27 15:33 /scripts/sudo/sudo1.6.tar -rw-r--r-1 root sys 46745600 Jul 27 09:48 /scripts/scripts.tar -rw-r--r-1 root system 10065920 Apr 20 2001 /scripts/exe/exe_files.tar

These Search Results are Stored in ==> /tmp/largefiles.out

Search Complete...EXITING...

Listing 18.2 findlarge.ksh shell script in action. (continued)

The output in Listing 18.2 is a listing of the entire screen output, which is also the contents of the $OUTFILE. The user is informed of the trigger threshold for the search, the top-level directory for the search, the hostname of the machine, the date and time of the search, and the number of files found to exceed the threshold. The long listing of each file is displayed that has the file owner and group, the size of the file in bytes, the modification time, and the full path to the file. The long listing is very helpful in large shops with thousands of users!

Other Options to Consider
The findlarge.ksh shell script is simple and does all of the basics for the system reporting, but it can be improved and customized for your particular needs. I think you will be interested in the following ideas: 1. The first thing you probably noticed is that the script uses the current directory as the top-level directory for the search path. You may want to add a second command-line argument so that you can specify a search path other than the current directory. You could add this user-supplied search path as an option, and if a search path is omitted you use the current directory to start the search. This adds a little more flexibility to the shell script. 2. Each time we run the findlarge.ksh shell script, we overwrite the $OUTFILE. You may, however, want to keep a month’s worth of reports on the system. An easy way to keep one month of reports is to use the date command and extract the day of the month, and then add this value as a suffix to the $OUTFILE file name definition. The following command will work:
OUTFILE=”/tmp/largefiles.out.$(date +%d)”

Over time our script will result in filenames largefile.out.01 through largefiles.out.31.

Finding “Large” Files
3. When searching large filesystems the search may take a very long time to complete. To give the user feedback that the search process is continuing you may want to add one of the progress indicators studied in Chapter 4. Two of the studied progress indicators would be appropriate, the rotating line and the series of dots. Look in Chapter 4 for details. 4. When we specify our search value we are just adding six zeros to the usersupplied integer value. But we are back to a basic question: Is one MB equal to 1,000,000 or 1,024,000? Because a System Administrator may not be the one reading the report, maybe a manager, I used the mathematical 1,000,000 and not the system-reported power-of-2 value. This is really a toss-up, so you make the decision on the value you want to use. The value is easy to change by doing a little math to multiply the user-supplied value by 1,024,000. 5. If you need to look for newly created files when a filesystem has just filled up, you can add the following command as a cross reference to find the true cause of the filesystem filling up: find $SEARCH_PATH -mtime 1 -print

473

This command will find all files that have been modified, or created, in the last 24 hours. You can redirect this output to a file and do a cross-reference to discover the files, and users, that actually caused the filesystem to fill up.

Summary
In this chapter we have shown how to search the system for large files and create a machine-specific report. As stated in the previous section, there are many ways to do the same task, and as always we have other options to consider. This chapter, along with filesystem monitoring in Chapter 5, can help keep filesystem surprises to a minimum. In the next chapter we are going to study techniques to capture a user’s keystrokes. Capturing keystrokes has many uses, from giving you an audit trail of all root access to keeping track of a problem contractor or user. I use this technique to keep an audit trail of all root access to the systems. I hope you gained some knowledge in this chapter, and I will see you in the next chapter!

Free & Share & Open

CHAPTER

19
Monitoring and Auditing User Key Strokes

In most large shops there is a need, at least occasionally, to monitor a user’s actions. You may even want to audit the keystrokes of anyone with root access to the system or other administration type accounts, such as oracle. Contractors on site can pose a particular security risk. Typically when a new application comes into the environment one or two contractors are on site for a period of time for installation, troubleshooting, and training personnel on the product. I always set up contractors in sudo (see Chapter 14 for more details on sudo) to access the new application account, after I change the password. sudo tracks only the commands that were entered with a date/time stamp. The detail of the command output from stdout and stderr does not get logged so you do not have a complete audit trail of exactly what happened if a problem arises. To get around this dilemma you can track a user’s keystrokes from the time he or she accesses a user account until the time he or she exits the account, if you have the space for the log file. This little feat is accomplished using the script command. The idea is to use sudo to kick off a shell script that starts a script session. When the script session is running, all of the input and output on the terminal is captured in the log file. Of course, if the user goes into some menus or programs the log file gets a little hard to read, but we at least have an idea what happened. This monitoring is not done surreptitiously because I always want everyone to know that the monitoring is taking place. When a script session starts, output from the script command informs the user that a session is running and gives the name of the session’s log file. We can also set up mon-

475

Free & Share & Open

476

Chapter 19 itoring to take place from the time a user logs in until the user logs out. For this monitoring we do not need sudo, but we do need to edit the $HOME/.profile or other login configuration file for the particular user.

Syntax
Using the script command is straightforward, but we want to do a few more things in the shell script. Giving a specific command prompt is one option. If you are auditing root access you need to have a timeout set so that after about five minutes (see the TMOUT environment variable) the shell times out and the root access ends. On a shell timeout, the session is terminated and the user is either logged out or presented with a command prompt, but we can control this behavior. We have many options for this set of shell scripts. You are going to need to set up sudo, super-user-do, on your machine. The full details for installing and configuring sudo are in Chapter 14. We want sudo to be configured with the names of each of the shell scripts that are used for this monitoring effort, as well as the specific users that you will allow to execute them. We will get to these details later. The script command works by making a typescript of everything that appears on the terminal. The script command is followed by a filename that will contain the captured typescript. If no filename is given the typescript is saved in the current directory in a file called typescript. For our scripting we will specify a filename to use. The script session ends when the forked shell is exited, which means that there are two exits required to completely log out of the system. The script command has the following syntax: script [filename]

As the script session starts, notification is shown on the terminal and a time stamp is placed at the top of the file, indicating the start time of the session. Let’s look at a short script session as used on the command line in Listing 19.1.
[root:yogi]@/# more /usr/local/logs/script/script_example.out Script command is started on Wed May 8 21:35:27 EDT 2002. [root:yogi]@/# cd /usr/spool/cron/crontabs [root:yogi]@/usr/spool/cron/crontabs# ls adm root sys uucp [root:yogi]@/usr/spool/cron/crontabs# ls -al total 13 drwxrwx--2 bin cron 512 Feb 10 21:36 . drwxr-xr-x 4 bin cron 512 Jul 26 2001 .. -rw-r--r-1 adm cron 2027 Feb 10 21:36 adm -rw------1 root cron 1125 Feb 10 21:35 root -rw-r--r-1 sys cron 864 Jul 26 2001 sys

Listing 19.1 Command-line script session.

Monitoring and Auditing User Key Strokes

477

-rw-r--r-1 root cron 703 Jul [root:yogi]@/usr/spool/cron/crontabs# cd ../.. [root:yogi]@/usr/spool# ls -l total 12 drwxrwsrwt 2 daemon staff 512 Sep drwxr-xr-x 4 bin cron 512 Jul drwxrwxr-x 7 lp lp 512 Mar drwxrwxr-x 5 bin printq 512 May drwxrwxr-x 2 bin mail 512 May drwxrwx--2 root system 512 May drwxrwxr-x 2 bin printq 512 Apr drwxr-xr-x 2 root system 512 Jul drwxrwsrwx 2 bin staff 512 Jul drwxr-xr-x 11 uucp uucp 512 Mar drwxrwxrwx 2 uucp uucp 512 Sep drwxrwxr-x 2 root system 512 Apr [root:yogi]@/usr/spool# exit Script command is complete on Wed May [root:yogi]@/#

26 2001

uucp

17 26 23 01 06 06 29 26 26 13 08 16

2000 2001 15:21 20:32 17:36 17:36 11:52 2001 2001 20:43 2000 2001

calendar cron lp lpd mail mqueue qdaemon rwho secretmail uucp uucppublic writesrv

8 21:36:11 EDT 2002.

Listing 19.1 Command-line script session. (continued)

Notice that every keystroke is logged as well as all of the command output. At the beginning and end of the log file a script command time stamp is produced. These lines of text are also displayed on the screen as the script session starts and stops. These are the user notifications given as the monitoring starts and stops.

Scripting the Solution
There are three different situations in which you want to use this type of monitoring/auditing. In this first instance we have users that you want to monitor the entire session. In the next situation you want to monitor activity only when a user wants root access to the system. Our systems have direct, remote, and su root login disabled, so to gain root access the user must use sudo to switch to root using the broot script. The third script is a catch-all for other administrative user accounts that you want to audit. The first script is covering end-to-end monitoring with the script execution starting at login through the user’s $HOME/.profile. Before we actually start the script session, there are some options to consider. Because we are executing a shell script from the user’s .profile we need to ensure that the script is the last entry in the file. If you do not want the users to edit any .profile files, then you need to set the ownership of the file to root and set the user to read-only access.

Free & Share & Open

478

Chapter 19

Logging User Activity
We are keeping log files so it is a good idea to have some kind of standard format for the log filenames. You have a lot of options for filenames, but I like to keep it simple. Our log files use the following naming convention:
[hostname].[user $LOGNAME].[Time Stamp]

We want the hostname because most likely you are monitoring users on multiple systems and using a central repository to hold all of the log files. When I write a shell script I do not want to execute a command more times than necessary. The hostname command is a good example. Assigning the system’s hostname to a variable is a good idea because it is not going to change, or it should not change, during the execution of the script. To assign the hostname of the system to a variable use the following syntax:
THISHOST=$(hostname)

For the date/time stamp a simple integer representation is best. The following date command gives two digits for month, day, year, hour, minute, and second:
TS=$(date +%m%d%y%H%M%S)

Now we have to reference only the $TS variable for the date/time stamp. Because the user may change we can find the active username with either of the following environment variables: echo $LOGNAME echo $USER echo $LOGIN

As you change user IDs by using the switch user command (su), all of these environment variables change accordingly. However, if a user does a switch user using sudo, then the $LOGIN environment variable carries over to the new user while the $LOGNAME and $USER environment variables gain the new user ID. Now we have everything to build a log filename. A good variable name for a log file is LOGFILE, unless this variable is used by your system or another application. On my systems the LOGFILE variable is not used. Not only do we need to create the name of the $LOGFILE, but we need to create the file and set the permissions on the file. The initial permissions on the file need to be set to read/write by the owner, chmod 600 $LOGFILE. The following commands set up the log file:
TS=$(date +%m%d%y%H%M%S) THISHOST=$(hostname) LOGFILE=${THISHOST}.${LOGNAME}.$TS touch ${LOGDIR}/$LOGFILE chmod 600 ${LOGDIR}/${LOGFILE} # # # # # Create a time stamp Query the system for the hostname Name the log file Create an empty log file Set the file permissions

Monitoring and Auditing User Key Strokes
A sample filename is shown here: yogi.randy.05110274519 479

The filename is good, but where do we want to store the file on the system? I like to use a separate variable to hold the directory name. With two separate variables representing the directory and filename, you can move the log directory to another location and have to change just one entry in the script. I set up a log directory on my system in /usr/local/logs. For these script log files I added a subdirectory called script. Then I set a LOGDIR variable to point to my logging directory, as shown here:
LOGDIR=/usr/local/logs/script

Starting the Monitoring Session
With the logging set up we are ready to start a script session. We start the session using the following syntax: script ${LOGDIR}/${LOGFILE}

When the script session starts, a message is displayed on the screen that informs the user that a script session has started and lists the name of the script log file, as shown here:
Script command is started. The file is /usr/local/logs/script/yogi.randy.051102174519.

If the user knows that monitoring is going on and also knows the name of the file, what is to keep the user from editing or deleting the log? Usually directory permissions will take care of this little problem. During the script session the actual log file is an open file—that is, actually a system temporary file that cannot be accessed directly by the user. But if the user is able to delete the $LOGFILE then you have lost the audit trail. This is one problem that we will discuss later.

Where Is the Repository?
So far here is the scenario. A user has logged into the system. As the user logs in, a monitoring session is started using the script command, which logs all of the terminal output in a log file that we specify. During the time that the session is active the log file is open as a system temporary file. When the session ends, by a user typing exit or CTRL-D or by an exit signal, the log file is closed and the user is notified of the session ending, and again the name of the log file is displayed. For security and auditing purposes we need to have a central repository for the logs. The method I like to use is email. When the session ends we want to set the file permissions on the log file to read only by the owner. Then we email the log to another machine, ideally, which is where the repository is located. Once the email is sent I compress the local file and exit the script.

Free & Share & Open

480

Chapter 19
With two copies of the user session existing on two different machines, an audit will easily detect any changes. In fact, if a user tries to change the log these commands will also be logged. You may have different ideas on handling the repository, but I set up a user on a remote machine that I use as a log file manager, with a name logman. The logman user’s email is the repository on the audit machine. For simplicity in this shell script we are going to email the logs to the local logman user. To send mail, I use the mailx command on all Unix flavors except Linux, where I use the mail command, as shown here: mailx -s “$TS - $LOGNAME Audit Report” $LOG_MANAGER < ${LOGDIR}/${LOGFILE}

In the shell script the $LOG_MANAGER is defined as logman. The nice thing about having a variable hold the mail recipients is that you can add a second repository or other people to receive email notifications. By using the local logman account you have other options. You can set up mail aliases; one of my favorites is to use the logman account as a bounce account. By adding a .forward file in the $HOME directory for the logman user, you can redirect all of the email sent to the logman user to other destinations. If a .forward file exists in the user’s home directory, the mail is not delivered to the user but instead is sent to each email address and alias listed in the .forward file. A sample .forward file is shown here. yogibear@cave.com booboo@cave.com dino@flintstones.org admin

With the previous entries in the $HOME/.forward file for the logman user, all mail directed to logman is instead sent to the three email address and all of the addresses pointed to by the admin email alias.

The Scripts
We have covered all of the basics for the shell scripts. We have three different shell scripts that are used in different ways. The first script is intended to be executed at login time by being the last entry in the user’s $HOME/.profile. The second shell script is used only when you want to gain root access, which is done through sudo, and the third script is a catch-all for any other administration-type accounts that you want to audit, which also use sudo. Let’s first look at the login script called log_keystrokes.ksh, shown in Listing 19.2.
#!/bin/ksh # # SCRIPT: log_keystrokes.ksh # # AUTHOR: Randy Michael

Listing 19.2 log_keystrokes.ksh shell script listing.

Monitoring and Auditing User Key Strokes

481

# DATE: 05/08/2002 # REV: 1.0.P # PLATFOEM: Any Unix # # PURPOSE: This shell script is used to monitor a login session by # capturing all of the terminal data in a log file using # the script command. This shell script name should be # the last entry in the user’s $HOME/.profile. The log file # is both kept locally and emailed to a log file # administrative user either locally or on a remote machine. # # REV LIST: # # # set -n # Uncomment to check syntax without any execution # set -x # Uncomment to debug this shell script # ############# DEFINE AUDIT LOG MANAGER ################### # # This user receives all of the audit logs by email. This # Log Manager can have a local or remote email address. You # can add more than one email address if you want by separating # each address with a space. LOG_MANAGER=”logman” # List to email audit log

########################################################## ################ DEFINE FUNCTIONS HERE ################### ########################################################## cleanup_exit () { # This function is executed on any type of exit except of course # a kill -9, which cannot be trapped. The script log file is # emailed either locally or remotely, and the log file is # compressed. The last “exit” is needed so the user does not # have the ability to get to the command line without logging. if [[ -s ${LOGDIR}/${LOGFILE} ]] then case `uname` in Linux) # Linux does not have “mailx” mail -s “$TS - $LOGNAME Audit Report” $LOG_MANAGER < ${LOGDIR}/${LOGFILE} ;; *) mailx -s “$TS - $LOGNAME Audit Report” $LOG_MANAGER < ${LOGDIR}/${LOGFILE} ;;

Listing 19.2 log_keystrokes.ksh shell script listing. (continues)

Free & Share & Open

482

Chapter 19

esac compress ${LOGDIR}/${LOGFILE} 2>/dev/null fi exit } # Set a trap trap ‘cleanup_exit’1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 26 ########################################################## ################ DEFINE VARIABLES HERE ################### ########################################################## TS=$(date +%m%d%y%H%M%S) THISHOST=$(hostname|cut -f1-2 -d.) LOGDIR=/usr/local/logs/script LOGFILE=${THISHOST}.${LOGNAME}.$TS touch $LOGDIR/$LOGFILE set -o vi 2>/dev/null stty erase ^? # # # # # # # File time stamp Host name of this machine Directory to hold the logs Creates the name of the log file Creates the actual file Previous commands recall Set the backspace key

# Set the command prompt export PS1=”[$LOGNAME:$THISHOST]@”’$PWD> ‘ #################### RUN IT HERE ########################## chmod 600 ${LOGDIR}/${LOGFILE} script ${LOGDIR}/${LOGFILE} chmod 400 ${LOGDIR}/${LOGFILE} # Change permission to RW for the owner # Start the script monitoring session # Set permission to read-only for # the owner # Execute the cleanup and exit function

cleanup_exit

Listing 19.2 log_keystrokes.ksh shell script listing. (continued)

The log_keystrokes.ksh script in Listing 19.2 is not difficult when you look at it. At the top we define the cleanup_exit function that is used when the script exits to email and compress the log file. In the next section we set a trap and define and set some variables. Finally we start the logging activity with a script session. In the cleanup_exit function notice the list of exit codes that the trap command will exit on. This signal list ensures that the log file gets emailed and the file gets compressed.

Monitoring and Auditing User Key Strokes
The only exit signal we cannot do anything about is a kill -9 signal because you cannot trap kill -9. There are more exit signals if you want to add more to the list in the trap statement, but I think the most captured are listed. The last command executed in this shell script is exit because in every case the cleanup_exit function must execute. If exit is not the last command, then the user will be placed back to a command prompt without any logging being done. The reason for this behavior is that the script session is really a fork of the original shell. Therefore, when the script command stops executing, one of the shells in the fork terminates, but not the original shell. This last exit logs out of the original shell. You may want to replace this last exit, located in the cleanup_exit function, with logout, which will guarantee the user is logged out of the system.

483

Logging root Activity
In some shops there is a need to log the activity of the root user. If you log the root activity, then you have an audit trail, and it is much easier to do root cause analysis on a root user booboo. We can use the same type of shell that we used in the previous sections, but this time we will use sudo instead of a .profile entry. I call this script broot because it is a short name for “I want to be root”. In this section let’s look at the shell script in Listing 19.3 and go through the details at the end.
#!/bin/ksh # # SCRIPT: broot # # AUTHOR: Randy Michael # DATE: 05/08/2002 # REV: 1.0.P # PLATFOEM: Any Unix # # PURPOSE: This shell script is used to monitor all root access by # capturing all of the terminal data in a log file using # the script command. This shell script is executed from the # command line using sudo (Super User Do). The log file # is kept locally and emailed to a log file administrative # user either locally or on a remote machine. Sudo must be # configured for this shell script. Refer to your sudo notes. # # USAGE: sudo broot # # REV LIST: # # # set -n # Uncomment to check syntax without any execution # set -x # Uncomment to debug this shell script #

Listing 19.3 broot shell script listing. (continues)

Free & Share & Open

484

Chapter 19

############# DEFINE AUDIT LOG MANAGER ################### # # # # This user receives all of the audit logs by email. This Log Manager can have a local or remote email address. You can add more than one email address if you want by separating each address with a space. # List to email audit log

LOG_MANAGER=”logman”

########################################################## ################ DEFINE FUNCTIONS HERE ################### ########################################################## cleanup_exit () { # This function is executed on any type of exit except of course # a kill -9, which cannot be trapped. The script log file is # emailed either locally or remotely, and the log file is # compressed. The last “exit” is needed so the user does not # have the ability to get to the command line without logging. if [[ -s ${LOGDIR}/${LOGFILE} ]] then case `uname` in Linux) # Linux does not have “mailx” mail -s “$TS - $LOGNAME Audit Report” $LOG_MANAGER < ${LOGDIR}/${LOGFILE} ;; *) mailx -s “$TS - $LOGNAME Audit Report” $LOG_MANAGER < ${LOGDIR}/${LOGFILE} ;; esac nohup compress ${LOGDIR}/${LOGFILE} 2>/dev/null & fi exit } # Set a trap trap ‘cleanup_exit’1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 26 ########################################################## ################ DEFINE VARIABLES HERE ################### ########################################################## TS=$(date +%m%d%y%H%M%S) # File time stamp

Listing 19.3 broot shell script listing. (continued)

Monitoring and Auditing User Key Strokes

485

THISHOST=$(hostname) LOGDIR=/usr/local/logs/script LOGFILE=${THISHOST}.${LOGNAME}.$TS touch $LOGDIR/$LOGFILE TMOUT=300 export TMOUT set -o vi stty erase _ # Run root’s .profile if one exists if [[ -f $HOME/.profile ]] then . $HOME/.profile fi

# # # # # # # #

Host name of this machine Directory to hold the logs Creates the name of the log file Creates the actual file Set the root shell timeout!!! Export the TMOUT variable To recall previous commands Set the backspace key

# set path to include /usr/local/bin echo $PATH|grep -q ‘:/usr/local/bin’ || PATH=$PATH:/usr/local/bin # Set the command prompt to override the /.profile default prompt PS1=”$THISHOST:broot> “ export PS1 #################### RUN IT HERE ##########################

chmod 600 ${LOGDIR}/${LOGFILE} script ${LOGDIR}/${LOGFILE} chmod 400 ${LOGDIR}/${LOGFILE} owner cleanup_exit

# Change permission to RW for the owner # Start the script monitoring session # Set permission to read-only for the

# Execute the cleanup and exit function

Listing 19.3 broot shell script listing. (continued)

There is one extremely important difference between this script and the script in Listing 19.2. In the broot script in Listing 19.3 we execute the .profile for root, if there is a .profile for root. You may ask why we did not execute the profile last time. The answer involves the recursive nature of running a file onto itself. In the previous case we had the following entry in the $HOME/.profile file:
. /usr/local/bin/log_keystrokes.ksh

We add this entry beginning with a “dot”, which means to execute the following file, as the last entry in the $HOME/.profile. If you added execution of $HOME/ .profile into the shell script you end up executing the log_keystrokes.ksh shell

Free & Share & Open

486

Chapter 19 script recursively. When you run the script like this you fill up the buffers and you get an error message similar to the following output: ksh: .: 0403-059 There cannot be more than 9 levels of recursion.

For monitoring root access with the broot script we are not executing from the .profile, but we use sudo to run this broot script, so we have no worries about recursion. At the top of the script in Listing 19.3 we define a LOG_MANAGER. This list of one or more email addresses is where the log files are going to be emailed. You may even want real-time notification of root activity. I like to send the log files off to my audit box for safe keeping using my logman user account. This email notice in the cleanup_exit function uses two different e-mail commands, depending on the Unix flavor. The only machine that does not support the mailx command is Linux, which supports only the mail command. This is not a problem, but I had to use the mix email commands to add a subject heading in the email; not all mail commands on all systems allow a subject heading so I used mailx instead. The next step is to set a trap. If the script exits on signals 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 26, the cleanup_exit function is executed. This trap ensures that the log file gets emailed and the file gets compressed locally. In the next section we define and set the variables that we use. Notice that we added a shell timeout, specified by the TMOUT environment variable. If someone with root access is not typing for five minutes the shell times out. You can set the TMOUT variable to anything you want or even comment it out if you do not want a shell timeout. The measurement is in seconds. The default is 300 seconds, or 5 minutes, for this script. After the variable definitions we execute the root .profile. We run the profile here because we are not running the broot script from a login $HOME/.profile, as we did with the log_keystrokes.ksh script in Listing 19.2. Next we add /usr/ local/bin to root’s $PATH, if it is not already present. And, finally, before we are ready to execute the script command we set a command prompt. The final four things we do are (1) set the permissions on the log file so we can write to it; (2) run the script command using the log filename as a parameter; (3) set the file permissions on the log file to read-only; and (4) execute the cleanup_exit function to email the log and compress the file locally.

Some sudo Stuff
I have inserted a short /etc/sudoers file for Listing 19.4 to show entries that need to be made. The entire task of setting up and using sudo is shown in Chapter 14. Pay attention to the bold type in Listing 19.4.
# sudoers file. # # This file MUST be edited with the ‘visudo’ command as root. # # See the sudoers man page for the details on how to write a

Listing 19.4 Example /etc/sudoers file.

Monitoring and Auditing User Key Strokes

487

# # # # # # # # # # # # # # # #

sudoers file.

Users Identification: All access: randy - Randy Michael terry - Admin Restricted Access to: mount umount and exportfs oracle - Oracle Admin operator - operator access Host alias specification LOCAL=yogi

Host_Alias

# User alias specification User_Alias User_Alias User_Alias User_Alias User_Alias User_Alias ROOTADMIN=randy,terry NORMAL=randy,operator,terry ADMIN=randy,terry ORACLE=oracle DB2=db2adm OPERATOR=operator

# Runas alias specification Runas_Alias ORA=oracle

# Cmnd alias specification Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias Cmnd_Alias BROOT=/usr/local/bin/broot MNT=/usr/bin/mount UMNT=/usr/bin/umount EXP_FS=/usr/bin/exportfs KILL=/usr/bin/kill ORACLE_SU=/usr/bin/su - oracle TCPDUMP=/usr/sbin/tcpdump ERRPT=/usr/bin/errpt SVRMGRL=/oracle/product/8.0.5/bin/svrmgrl

# User privilege specification root ALL=(ALL) ALL ROOTADMIN LOCAL=BROOT NORMAL LOCAL=MNT,UMNT,EXP_FS

Listing 19.4 Example /etc/sudoers file. (continues)

Free & Share & Open

488

Chapter 19

ADMIN LOCAL=BROOT,MNT,UMNT,KILL,ORACLE_SU,TCPDUMP,ERRPT: \ LOCAL=EXP_FS ORACLE LOCAL=SVRMGRL # Override Defaults Defaults logfile=/var/adm/sudo.log

Listing 19.4 Example /etc/sudoers file. (continued)

Three entries need to be added to the /etc/sudoers file. Do not ever edit the sudoers file directly with vi. There is a special program called visudo, in the /usr/ local/sbin directory, that has a wrapper around the vi editor that does a thorough check for mistakes in the file before the file is saved. If you make a mistake the visudo program will tell you where the error is located in the /etc/sudoers file. The three entries that need to be added to the /etc/sudoers are listed next and are highlighted in bold text in Listing 19.4. Define the User_Alias, which is where you give a name to a group of users. For this file let’s name the list of users who can get root access ROOTADMIN, as shown here:
User_Alias ROOTADMIN=randy,terry

Next we need to define the Cmnd_Alias, which is where you define the full pathname to the command, as shown here.
Cmnd_Alias BROOT=/usr/local/bin/broot

The last step is to define the exact commands that the User_Alias group of users can execute. In our case we have a separate User_Alias group only for the users who can use the broot script. Notice that the definition also specifies the machine where the command can be executed. I always let sudo execution take place only on a single machine at a time, specified by LOCAL here.
ROOTADMIN LOCAL=BROOT

Once the /etc/sudoers file is set up, you can change the root password and allow root access only by using the broot script. Using this method you have an audit trail of root access to the system.

Monitoring and Auditing User Key Strokes

489

Monitoring Other Administration Users
More often than not, you will want add to the list of auditing that can be done. This next script is rewritten to allow you to quickly set up a broot type shell script by changing only the user name and the script name. The method that we use to execute the script command is what makes this script different—and easy to modify. For ease of use we can use a lot of variables throughout the script. We have already been doing this to some extent. Now we will call the monitored user the effective user, which fits our new variable $EFF_USER. For this script I have set the username to oracle. You can make it any user that you want to. Take a look at this shell script in Listing 19.5, and pay particular attention to the boldface type.
#!/bin/ksh # # SCRIPT: “Banybody” boracle - This time # # AUTHOR: Randy Michael # DATE: 05/08/2002 # REV: 1.0.P # PLATFOEM: Any Unix # # PURPOSE: This shell script is used to capture all “$EFF_USER” # access by capturing all of the terminal data in a log # file using the script command. This shell script is # executed from the command line using sudo (Super User Do). # The log file is kept locally and emailed to a log file # administrative user either locally or on a remote # machine. Sudo must be configured for this shell script. # Refer to your sudo notes. The effective user, currently # oracle, can be changed by setting the “EFF_USER” variable # to another user, and changing the name of the script. # This is why the original name of the script is called # “Banybody”. # # ORIGINAL USAGE: sudo Banybody # # THIS TIME USAGE ==> USAGE: sudo boracle # # # REV LIST: # 5/10/2002: Modified the script to replace the hard-coded # username with the variable $EFF_USER. This # allows flexibility to add auditing of more

Listing 19.5 boracle shell script listing. (continues)

Free & Share & Open

490

Chapter 19

# accounts by just changing the EFF_USER variable # and the script name. # # set -n # Uncomment to check syntax without any execution # set -x # Uncomment to debug this shell script # # ################# DEFINE EFFECTIVE USER ################## # This EFF_USER is the username you want to be to execute # a shell in. An su command is used to switch to this user. EFF_USER=oracle ############# DEFINE AUDIT LOG MANAGER ################### # # # # This user receives all of the audit logs by email. This Log Manager can have a local or remote email address. You can add more than one email address if you want by separating each address with a space. # List to email audit log

LOG_MANAGER=”logman”

########################################################## ################ DEFINE FUNCTIONS HERE ################### ########################################################## cleanup_exit () { # This function is executed on any type of exit except of course # a kill -9, which cannot be trapped. The script log file is # emailed either locally or remotely, and the log file is # compressed. The last “exit” is needed so that the user does not # have the ability to get to the command line without logging. if [[ -s ${LOGDIR}/${LOGFILE} ]] # Is it greater than zero bytes? then case `uname` in Linux) mail -s “$TS - $LOGNAME Audit Report” $LOG_MANAGER < ${LOGDIR}/${LOGFILE} ;; *) mailx -s “$TS - $LOGNAME Audit Report” $LOG_MANAGER < ${LOGDIR}/${LOGFILE}

Listing 19.5 boracle shell script listing. (continued)

Monitoring and Auditing User Key Strokes

491

;; esac compress ${LOGDIR}/${LOGFILE} 2>/dev/null fi exit } ################# SET A TRAP ############################# trap ‘cleanup_exit’ 1 2 3 5 15 ########################################################## ################ DEFINE VARIABLES HERE ################### ########################################################## TS=$(date +%m%d%y%H%M%S) THISHOST=$(hostname) LOGDIR=/usr/local/logs/script LOGFILE=${THISHOST}.${EFF_USER}.$TS touch $LOGDIR/$LOGFILE TMOUT=300 export TMOUT set -o vi stty erase ^? # # # # # # # # # File time stamp Hostname of this machine Directory to hold the logs Creates the name of the log file Creates the actual file Set the root shell timeout!!! Export the TMOUT variable To recall previous commands Set the backspace key

# set path to include /usr/local/bin echo $PATH|grep -q ‘:/usr/local/bin’ || PATH=$PATH:/usr/local/bin # Set the command prompt to override the /.profile default prompt PS1=”$THISHOST:b${EFF_USER}> “ export PS1 #################### RUN IT HERE ########################## chmod 666 ${LOGDIR}/${LOGFILE} # # # # # # # Set permission to read/write

To get the script session to work we have to use the switch user (su) command with the -c flag, which means execute what follows. Sudo is also used just to ensure that root is executing the su command. We ARE executing now as root because this script was started with sudo. If a nonconfigured sudo user tries to execute this command then it will fail unless sudo was used to execute this script as root.

Listing 19.5 boracle shell script listing. (continues)

Free & Share & Open

492

Chapter 19

# Notice we are executing the script command as “$EFF_USER”. This # variable is set at the top of the script. A value such as # “EFF_USER=oracle” is expected. sudo su - $EFF_USER -c “script ${LOGDIR}/${LOGFILE}” chmod 400 ${LOGDIR}/${LOGFILE} # Set permission to read-only for # the owner # Execute the cleanup and exit function

cleanup_exit

Listing 19.5 boracle shell script listing. (continued)

The most important line to study in Listing 19.5 is the third line from the bottom: sudo su - $EFF_USER -c “script ${LOGDIR}/${LOGFILE}”

There are several points to make about this command. Notice that we start the command with sudo. Because you must use sudo to execute the boracle script, and you are already executing as root, then why use sudo here? We use sudo here to ensure that the boracle script was indeed started with sudo. If any old user runs the boracle command we want it to fail if sudo was not used. The second command in the previous statement is su - $EFF_USER. The significance of the hyphen, -, is important here. Using the hyphen, -, with a space on both sides tells the su command to switch to the user pointed to by the $EFF_USER, oracle in our case, and run that user’s .profile. If the hyphen is omitted or the spaces are not around the hyphen, then the user .profile is not executed, which is a bad thing in this case. The last part of this command is where we start our script session. When you switch users with su, you can specify that you want to run a command as this user by adding the -c switch followed by the command enclosed in single or double quotes. Do not forget the quotes around the command. The only other real change is the use of the EFF_USER variable. This variable is set at the top of the script, and changing this variable changes who you want to “be.” If you want to create more admin auditing scripts, copy the boracle file to a new filename and edit the file to change the name at the top of the script and modify the EFF_USER variable. That’s it!

Other Options to Consider
Through this chapter we have covered some interesting concepts. You may have quite a few things that you want to add to these scripts. I have come up with a few myself.

Monitoring and Auditing User Key Strokes

493

Emailing the Audit Logs
Depending on the extent of monitoring and auditing you need to do, you may want to send the files to several different machines. I selected using email for the transport, but you may have some other techniques, such as automated FTP. You may also want to compress the files before you email, or whatever, the log files. To email a compressed file you will need some type of mail tool like metasend or a tool that does a type of uuencoding. This is needed sometimes because the mail program will think that some of the characters, or control characters, are mail commands. This can cause some strange things to happen. You should be able to find some mail tools on the Web. Watch the disk space! When you start logging user activity you need to keep a close check on disk space. Most systems store email in /var. If you fill up /var for an extended period of time you may crash the box. For my log files I create a large dedicated filesystem called /usr/local/logs. With a separate filesystem I do not have to worry about crashing the system if I fill up the filesystem. You can probably think of other methods to move the files around as the emails are received.

Compression
For all of these scripts we used the compress command. This compression algorithm is okay, but we can do better. I find that gzip has a much better compression algorithm, and the compression ratio is tunable for your needs. The tuning is done using numbers as a parameter to the gzip command, as shown here:
# gzip -9 $LOGFILE

The valid numbers are 1 to 9, with 9 indicating the best compression. This extra compression does come at a price—time! The higher the number, the longer it takes to compress the file. By omitting the number you use gzip in default mode, which is -5. For our needs you will still see a big increase in compression over compress at about the same amount of time.

Need Better Security?
Another option for this keystroke auditing is to use open secure shell and keep a real time encrypted connection to the log server by creating a named pipe. This can be done but it, too, has some potential problems. This first major problem is that you introduce a dependency for the logging to work. If the connection is lost then the script session ends. For auditing root activities, and especially when all other root access has been disabled, you can have a real nightmare. I will leave this idea for you to play around with because it is beyond the scope of this book.

Inform the Users
I did not add this chapter to the book for everyone to start secretly monitoring everyone’s keystrokes. Always be up-front with the user community, and let them know that

Free & Share & Open

494

Chapter 19 an audit is taking place. I know for a fact that Systems Administrators do not like to have the root password taken away from them. I know first hand about the reaction. If you are going to change the user password, please place the root password in a safe place where, in case of emergency, you can get to the password without delay. Your group will have to work out how this is accomplished.

Sudoers File
If you start running these scripts and you have a problem, first check your sudo configuration by looking at the /etc/sudoers file. There are some things to look for that the visudo editor will not catch:
■ ■

Check the LOCAL line. This variable should have the hostname of your machine assigned. Check for exact pathnames of the files. Ensure that the correct users are assigned to the correct commands.

■ ■ ■ ■

The visudo editor does catch most errors, but there are some things that are not so easy to test for.

Summary
I had a lot of fun writing this chapter and playing with these scripts. I hope you take these auditing scripts and use them in a constructive way. The information gathered can be immense if you do not have a mechanism for pruning the old log files. The following command works pretty well: find /directory -mtime +30 -print -exec rm {} \;

This command will remove all the files in /directory that have not been modified in 30 days. You may want to add a -name parameter to limit what you delete. As with any type of monitoring activity that creates logs, you need to watch the filesystem space very closely, especially at first, to see how quickly logs are being created and how large the log files grow. Another topic that comes up a lot is the shell timeout. The only place I use the TMOUT environment is in the broot script. If you add a shell timeout to your other administrative accounts you may find that a logout happens during a long processing job. With these users I expect them to just lock the terminal when they leave. In the next chapter we are going to look at Serial Storage Architecture (SSA) disk drives and how to physically identify them. These drives normally come in a rack of 18 drives, and we have a ton of racks! In this mess it is hard to locate a specific drive or a group of drives. We have a script that turns the identification lights on and off, with a lot of different options. See you in the next chapter!

CHAPTER

20
Turning On/Off SSA Identification Lights

On any system that utilizes the Serial Storage Architecture (SSA) disk subsystem from IBM you understand how difficult it is to find a specific failed disk in the hundreds of disks that populate the racks. Other needs for SSA disk identification include finding all of the drives attached to a particular system. Then you may also want to see only the drives that are in currently varied-on volume groups or a specific group of disks. In this chapter we will work through all of these areas of identification. In identifying hardware components in a system you usually have a set of tools for this function. This chapter is going to concentrate on AIX systems. The script presented in this chapter is valid only for AIX, but with a few modifications it can run on other Unix flavors that utilize the SSA subsystem. I am sticking to AIX because this script has an option to query volume groups, which not all Unix flavors support. If your systems are running the Veritas filesystem, then only a few commands need to be modified for my identification script to work because Veritas supports the concept of a volume group. In identifying an SSA disk you have two ways of referencing the disk. In AIX all disks are represented as an hdisk#. As an example, hdisk0 almost always contains the operating system, and it is part of the rootvg volume group. It is not often an SSA disk; it is usually an internal SCSI disk. If an hdisk is an SSA disk, then it has a second disk name that is used within the SSA subsystem, which is called the pdisk#. Not often are the hdisk# and the pdisk# the same number because the first couple of disks are usually SCSI drives. We need to be able to translate an hdisk to its associated pdisk, and vice versa.

495

Free & Share & Open

496

Chapter 20

Syntax
As always, we need to start out with the commands to accomplish the task. With the SSA subsystem we are concerned about two commands that relate to hdisks and pdisks. The first command, ssaxlate, translates an hdisk# into a pdisk#, or vice versa. The second command we use is the ssaidentify command, which requires a pdisk representation of the SSA disk drive. This command is used to turn the SSA disk identification lights on and off. We want the script to identify the SSA disks to recognize either disk format, hdisk and pdisk. With the ssaxlate command this is not a problem. To use these commands you need to know only the SSA disk to act on and add the appropriate command switch. Let’s look at both commands here.

Translating an hdisk to a pdisk
# ssaxlate -l hdisk43 pdisk41

In this example hdisk43 translates to pdisk41. This tells me that the hdisk to pdisk offset is 2, which I have to assume means that hdisk0 and hdisk1 are both SCSI disks, and hdisk3 through, at least, hdisk43 are all SSA disks. This is not always the case. It depends on how the AIX configuration manager discovered the disks in the first place, but my statement is a fair assumption. We could just as easily translate pdisk41 to hdisk43 by specifying pdisk41 in the ssaxlate command. The next step is to actually turn on the identification light for hdisk43, which we discovered to be pdisk41. The ssaidentify command wants the disks represented as pdisks, so we need to use pdisk41 for this command.

Identifying an SSA Disk
# ssaidentify -l pdisk41 -y

The ssaidentify command will just return a return code of success or failure, but no text is returned. If the return code is 0, zero, then the command was successful. If the return code is nonzero, then the command failed for some reason and a message is sent to standard error, which is file descriptor 2. All we are interested in is if the return code is zero or not.

Turning On/Off SSA Identification Lights
Table 20.1 SSA Identification Functions PURPOSE Shows the user how to use the shell script Shows detailed information on how to use the shell script Executes when a trapped exit signal is detected Used to give the user feedback that processing continues Controls SSA identification lights for all system SSA disks Controls SSA disks only in currently varied-on volume groups Controls SSA identification of a list of one or more disks

497

FUNCTION NAME usage man_page cleanup twirl all_defined_pdisks all_varied_on_pdisks list_of_disks

The Scripting Process
In the SSA identification script we are going to use a lot of functions. These functions perform the work so we just need the logic to decide which function to execute. An important thing you need to understand about functions is that the function must be declared, or written, in the code previous to when you want to execute the function. This makes sense if you think about it: You have to write the code before you can use it! The functions involved in this shell script are listed in Table 20.1 for your convenience.

Usage and User Feedback Functions
As you can see, we have our work cut out for us, so let’s get started. The first function is the usage function. When a user input error is detected you want to give the user some feedback on how to properly use the shell script. Always create a usage function. I want to show you this function because I did something you may not know that you can do. I used a single echo command and have 15 separate lines of output. Take a look at the function in Listing 20.1 to see the method.

Free & Share & Open

498

Chapter 20

function usage { echo “\nUSAGE ERROR... \nMAN PAGE ==> $SCRIPTNAME -? \nTo Turn ALL Lights Either ON or OFF: \nUSAGE: SSAidentify.ksh [-v] [on] [off] EXAMPLE: SSAidentify.ksh -v on \nWill turn ON ALL of the system’s currently VARIED ON SSA identify lights. NOTE: The default is all DEFINED SSA disks \nTo Turn SPECIFIC LIGHTS Either ON or OFF Using EITHER the pdisk#(s) AND/OR the hdisk#(s): \nUSAGE: SSAidentify.ksh [on] [off] pdisk{#1} [hdisk{#2}]... EXAMPLE: SSAidentify.ksh on hdisk36 pdisk44 pdisk47 \nWill turn ON the lights to all of the associated pdisk#(s) that hdisk36 translates to and PDISKS pdisk44 and pdisk47. \nNOTE: Can use all pdisks, all hdisks or BOTH hdisk and pdisk together if you want...” exit 1 }

Listing 20.1 Usage function with a single echo command.

As you can see in Listing 20.1, I enclose the entire text that I want to echo to the screen within double quotes, “usage text”. To place text on the next line, just press the ENTER key. If you want an extra blank line or a TAB, then use one or more of the many cursor functions available with the echo command, as shown in Table 20.2. There are many more in the man pages on your system. When incorrect usage of the shell script is detected, which you have to build in to the script, the proper usage message in Listing 20.2 is displayed on the screen.

Table 20.2

Cursor Control Commands for the echo Command PURPOSE Insert a new line with a carriage return Tab over on TAB length characters for each \t entered Back the cursor up one space for each \b entered Leaves the cursor at the current position, without a carriage return or line feed

ECHO FUNCTION \n \t \b \c

Turning On/Off SSA Identification Lights

499

USAGE ERROR... MAN PAGE ==> SSAidentify.ksh -?

To Turn ALL Lights Either ON or OFF: USAGE: SSAidentify.ksh [-v] [on] [off] EXAMPLE: SSAidentify.ksh -v on Will turn ON ALL of the system’s currently VARIED ON SSA identify lights. NOTE: The default is all DEFINED SSA disks To Turn SPECIFIC LIGHTS Either ON or OFF Using EITHER the pdisk#(s) AND/OR the hdisk#(s): USAGE: SSAidentify.ksh [on] [off] pdisk{#1} [hdisk{#2}]... EXAMPLE: SSAidentify.ksh on hdisk36 pdisk44 pdisk47 Will turn ON the lights to all of the associated pdisk#(s) that hdisk36 translates to and PDISKS pdisk44 and pdisk47. NOTE: Can use all pdisks, all hdisks or BOTH hdisk and pdisk together if you want...

Listing 20.2 Example of cursor control using the echo command.

By using cursor control with the echo command, we can eliminate using a separate echo command on every separate line of text we want to display. I do the same thing in the man_page function. You can see this function in its entirety in the full shell script shown in Listing 20.9. Before I show you the cleanup function, I want to show you the twirl function. The twirl function is used to give feedback to the user, which you saw back in Chapter 4. As a brief review, the twirl function displays the appearance of a line rotating. And this is accomplished through? You guess it, cursor control using the echo command. I like the twirl function because it is not too hard to understand and it is very short. This function works by starting an infinite while loop, which is done using the : (colon) no-op operator. A no-op does nothing and always has a zero return code so it is perfect to create an infinite loop. The next step is to have a counter that counts only from 0 to 4. When the counter reaches 4 it is reset back to 0, zero. At each count a case statement is used to decide which of the four lines, -, \, |, and /, is to be displayed. At the same time, the cursor is backed up so it is ready to overwrite the previous line character with a new one. There is a sleep for one second on each loop iteration. You must

Free & Share & Open

500

Chapter 20 leave the sleep statement in the code or you will see a big load on the system by all of the continuous updates to the screen. I use this function for giving user feedback when a time-consuming job is executing. When the job is finished I kill the twirl function and move on. The easiest way to kill a background function is to capture the PID just after kicking off the background job, which is assigned to the $! shell variable. This is similar to the way $? is used to see the return code of the last command. The twirl function is shown in Listing 20.3. function twirl { TCOUNT=”0”

# For each TCOUNT the line twirls one increment

while : # Loop forever...until you break out of the loop do TCOUNT=$(expr ${TCOUNT} + 1) # Increment the TCOUNT case ${TCOUNT} in “1”) echo ‘-’”\b\c” sleep 1 ;; “2”) echo ‘\\’”\b\c” sleep 1 ;; “3”) echo “|\b\c” sleep 1 ;; “4”) echo “/\b\c” sleep 1 ;; *) TCOUNT=”0” ;; # Reset the TCOUNT to “0”, zero. esac done # End of twirl function }

Listing 20.3 Twirl function listing.

When I have a time-consuming job starting, I start the twirl function with the following commands: twirl & TWIRL_PID=$!

This leads into the next function, cleanup. In normal operation the twirl function is killed in the main body of the script, or in the function that it is called in, by using the

Turning On/Off SSA Identification Lights kill command and the previously saved PID, which is pointed to by the TWIRL_PID variable. Life, though, is not always normal. In the top of the main body of the shell script we set a trap. The trap is used to execute one or more commands, programs, or shell scripts when a specified exit code is captured. Of course, you cannot trap a kill -9! In this shell script we execute the cleanup function on exit codes 1, 2 ,3, 5, and 15. You can add more exit codes if you want. This cleanup function displays a message on the screen that a trap has occurred and runs the kill -9 $TWIRL_PID command before exiting the shell script. If you omit the trap and the twirl function is running in the background, it will continue to run in the background! You cannot miss it—you always have a twirling line on your screen. Of course, you can kill the PID if you can find it in the process table with the ps command. The cleanup function is shown in Listing 20.4.

501

function cleanup { echo “\n...Exiting on a trapped signal...EXITING STAGE LEFT...\n” kill -9 $TWIRL_PID # End of cleanup function }

Listing 20.4 Cleanup function listing.

When an exit code is captured the user is informed that the shell script is exiting, and then the kill command is executed on the PID saved in the $TWIRL_PID variable.

Control Functions
Now we get into the real work of turning on and off the SSA identification lights starting with the all_defined_pdisks function. This function is the simplest of the SSA identification functions in this chapter. The goal is to get a list of every SSA disk on the system and use the pdisk# to control the identification lights by turning all lights on or off in sequence. To understand this function you need to understand an AIX command called lsdev and the switches we use to extract only the pdisk information. The lsdev command is used to display devices in the system and the characteristics of devices. The -C switch tells the lsdev command to look at only the currently defined devices. Then the -c command switch is added to specify the particular class of device; in our case the device class is pdisk. So far our lsdev command looks like the following statement:
# lsdev -Cc pdisk

Free & Share & Open

502

Chapter 20
But we want to drill down a little deeper in the system. We can also specify a subclass to the previously defined class by adding the -s switch with our subclass ssar. We also want to have a formatted output with column headers so we add the -H switch. These headers just help ensure that we have good separation between fields. Now we have the following command:
# lsdev -Cc pdisk -s ssar -H

Using this command on a system with SSA disks you see an output similar to the one in Listing 20.5. name pdisk0 pdisk1 pdisk2 pdisk3 pdisk4 pdisk5 pdisk6 pdisk7 pdisk8 pdisk9 pdisk10 pdisk11 pdisk12 pdisk13 pdisk14 pdisk15 status Available Available Available Available Available Available Available Available Available Available Available Available Available Available Available Available location 34-08-5B91-01-P 34-08-5B91-02-P 34-08-5B91-03-P 34-08-5B91-04-P 24-08-5B91-05-P 24-08-5B91-07-P 24-08-5B91-06-P 24-08-5B91-08-P 24-08-5B91-09-P 24-08-5B91-10-P 24-08-5B91-11-P 24-08-5B91-12-P 34-08-5B91-13-P 34-08-5B91-14-P 34-08-5B91-16-P 34-08-5B91-15-P description SSA160 SSA160 SSA160 SSA160 SSA160 SSA160 SSA160 SSA160 SSA160 SSA160 SSA160 SSA160 SSA160 SSA160 SSA160 SSA160 Physical Physical Physical Physical Physical Physical Physical Physical Physical Physical Physical Physical Physical Physical Physical Physical Disk Disk Disk Disk Disk Disk Disk Disk Disk Disk Disk Disk Disk Disk Disk Disk Drive Drive Drive Drive Drive Drive Drive Drive Drive Drive Drive Drive Drive Drive Drive Drive

Listing 20.5 lsdev listing of pdisks.

In Listing 20.5 we have more information than we need. The only part of this lsdev command output that we are interested in is in the first column, and only the lines that have “pdisk” in the first column. To filter this output we need to expand our lsdev command by adding awk and grep to filter the output. Our expanded command is shown here:
# lsdev -Cc pdisk -s ssar -H | awk ‘{print $1}’ | grep pdisk

In this command statement we extract the first column using the awk statement in a pipe, while specifying the first column with the ‘{print $1}’ notation. Then we use grep to extract only the lines that contain the pattern pdisk. The result is a list of all currently defined pdisks on the system.

Turning On/Off SSA Identification Lights
To control the identification lights for the pdisks in this list we use a for loop and use our lsdev command to create the list of pdisks with command substitution. These steps are shown in Listing 20.6. function all_defined_pdisks { # TURN ON/OFF ALL LIGHTS: # Loop through each of the system’s pdisks by using the # command with the “-Cc pdisk” switch while using “awk” # out the actual pdisk number. We will either # turn the identifier lights on or off specified by the # variable: # # Turn lights on: -y # Turn lights off: -n # # as the $SWITCH value to the “ssaidentify” command, as

503

“lsdev” to extract $SWITCH

used below...

echo “\nTurning $STATE ALL of the system’s pdisks...Please Wait...\n” for PDISK in $(lsdev -Cc pdisk -s ssar -H | awk ‘{print $1}’ \ | grep pdisk) do echo “Turning $STATE ==> $PDISK” ssaidentify -l $PDISK -${SWITCH} \ || echo “Turning $STATE $PDISK Failed” done echo “\n...TASK COMPLETE...\n” }

Listing 20.6 all_defined_pdisks function listing.

In Listing 20.6 notice the command substitution used in the for loop, which is in bold text. The command substitution produces the list arguments that are assigned to the $PDISK variable on each loop iteration. As each pdisk is assigned, the ssaidentify command is executed using the $PDISK definition as the target and uses the $SWITCH as the action to take, which can be either -y for light on or -n for light off. These values are defined in the main body of the shell script. As each light is being turned on or off the user is notified. If the action fails the user is notified of the failure also. This failure notification is done using a logical OR, specified by the double pipes, ||. The next function is all_varied_on_pdisks. This function is different in that we must approach the task of getting a list of SSA disks to act on using completely different strategy. The result we want is the ability to control the SSA disks that are in volume

Free & Share & Open

504

Chapter 20 groups that are currently varied-on. To get this list we must first get a list of the variedon volume groups using the lsvg -o command. This command gives a list of varied-on volume groups directly without any added text so we are okay with this command’s output. Using this list of volume groups we can now use the lspv command to get a full listing of defined hdisks. From this list we use grep to extract the hdisks that are in currently varied-on volume groups. Notice that all of this activity so far is at the hdisk level. We need to have pdisks to control the identification lights. To build a list of hdisks to convert we use a for loop tagging on the volume groups with the VG variable. For each $VG we run the following command to build a list.
# lspv | grep $VG >> $HDISKFILE

Notice that we use a file to store this list. A file is needed because if a variable were used we might exceed the character limit for a variable, which is 2048 on most systems. As you know, most large shops have systems with hundreds, if not thousands, of SSA disks. To be safe we use a file for storage here. Using this list of hdisks we are going to use another for loop to translate each of the hdisks into the associated pdisk. Because we may still have a huge list containing pdisks we again use a file to hold the list. The translation takes place using the ssaxlate command, but what if some of these hdisks are not SSA disks? Well, the translation will fail! To get around this little problem we first test each translation and send all of the output to the bit bucket and check the return code of the ssaxlate command. If the return code is 0, zero, then the hdisk is an SSA disk. If the return code is nonzero, then the hdisk is not an SSA disk. The result is that only pdisks are added to the new pdisk list file, which is pointed to by the PDISKFILE variable. Because this translation may take quite a while we start the twirl function, which is our progress indicator, in the background before the translation begins. As soon as the translation process ends, the twirl function is killed using the saved PID. The only thing left to do is to perform the desired action on each of the pdisk identification lights. We do this by starting yet another for loop. This time we use command substitution to produce a list of pdisks by listing the pdisk list file with the cat command. On each loop iteration the ssaidentify command is executed for each pdisk in the list file. The all_varied_on_pdisk function is shown in Listing 20.7. function all_varied_on_pdisks { trap ‘kill -9 $TWIRL_PID; return 1’ 1 2 3 15 cat /dev/null > $HDISKFILE cat /dev/null > $PDISKFILE echo “\nGathering a list of Varied on system SSA disks...Please wait...\c” VG_LIST=$(lsvg -o) # Get the list of Varied ON Volume Groups for VG in $(echo $VG_LIST)

Listing 20.7 all_varied_on_pdisks function listing.

Turning On/Off SSA Identification Lights

505

do lspv | grep $VG >> $HDISKFILE # List of Varied ON PVs done twirl & # Gives the user some feedback during long processing times... TWIRL_PID=$! echo “\nTranslating hdisk(s) into the associated pdisk(s) ...Please Wait...\c” for DISK in $(cat $HDISKFILE) # Translate hdisk# into pdisk#(s) do # Checking for an SSA disk /usr/sbin/ssaxlate -l $DISK # 2>/dev/null 1>/dev/null if (($? == 0)) then /usr/sbin/ssaxlate -l $DISK >> $PDISKFILE # Add to pdisk List fi done kill -9 $TWIRL_PID # Kill the user feedback function... echo “\b “ # Clean up the screen by overwriting the last character echo “\nTurning $STATE all VARIED ON system pdisks...Please Wait...\n” # Act on each pdisk individually... for PDISK in $(cat $PDISKFILE) do echo “Turning $STATE ==> $PDISK” /usr/sbin/ssaidentify -l $PDISK -${SWITCH} \ ||echo “Turning $STATE PDISK Failed” done echo “\n\t...TASK COMPLETE...\n” }

Listing 20.7 all_varied_on_pdisks function listing. (continued)

Notice that there is a trap at the beginning of this function in Listing 20.7. Because we are using the twirl function for user feedback we need a way to kill off the rotating line so we added a trap inside the function. In the next step we initialized both of the list files to empty files. Then the fun starts. This is where we filter through all of the hdisks to find the ones that are in currently varied-on volume groups. With this hdisk list we loop through each of the disks looking for SSA disks. As we find each hdisk it is translated into a pdisk and added to the pdisk list. With all of the pdisks of interest found we loop through each one and turn on/off the SSA identification lights.

Free & Share & Open

506

Chapter 20
The last function is list_of_disks, which acts on one or more hdisks or pdisks that are specified on the command line when the shell script is executed. In the main body of the shell script we do all of the parsing of the command-line arguments because if you tried to parse the command line inside a function the parsing would act on the function’s argument, not the shell script’s arguments. Therefore this is a short function. In the main body of the shell script a variable, PDISKLIST, is populated with a list of pdisks. Because the user can specify either hdisks or pdisks, or both, on the command line the only verification that has been done is on the hdisks only, when they were translated to pdisks. We need do a sanity check to make sure that each of the pdisks we act on has a character special file in the /dev filesystem. This is done using the -c switch in an if...then test. If the pdisk listed has a character special file associated with it, then an attempt is made to turn the SSA identification light on/off, otherwise, the user is notified that the specified pdisk is not defined on the system. The list_of_disks function is shown in Listing 20.8. function list_of_disks { # TURN ON/OFF INDIVDUAL LIGHTS: # Loop through each of the disks that was passed to this script # via the positional parameters greater than $1, i.e., $2, $3, $4... # We first determine if each of the parameters is a pdisk or an # hdisk. For each hdisk passed to the script we first need to # translate the hdisk definition into a pdisk definition. This # script has been set up to accept a combinition of hdisks and # pdisks. # # We will either turn the identifier lights on or off specified by # the $SWITCH variable for each pdisk#: # # Turn lights on: -y # Turn lights off: -n # # as the $SWITCH value to the “ssaidentify” command, as used below... echo “\n” # The disks passed to this script can be all hdisks, all pdisks, # or a combination of pdisks and hdisks; it just does not matter. # We translate each hdisk into the associated pdisk(s).

echo “\nTurning $STATE individual SSA disk lights...\n” for PDISK in $(echo $PDISKLIST) do

Listing 20.8 list_of_disks function listing.

Turning On/Off SSA Identification Lights

507

# Is it a real pdisk?? if [ -c /dev/${PDISK} ] 2>/dev/null then # Yep - act on it... /usr/sbin/ssaidentify -l $PDISK -${SWITCH} >/dev/null if (($? == 0)) then /usr/bin/ssaxlate -l $PDISK -${SWITCH} if (($? == 0)) then echo “Light on $PDISK is $STATE” else echo “Turning $STATE $PDISK Failed” fi fi else else echo “\nERROR: $PDISK is not a defined device on $THISHOST\n” fi done echo “\n...TASK COMPLETE...\n” }

Listing 20.8 list_of_disks function listing. (continued)

Notice in the boldface text in Listing 20.8 where we do the test to see if the pdisk listed is a real pdisk by using the -c switch in the if statement. We have covered the rest of the function, so let’s move on to the main body of the shell script.

The Full Shell Script
This is a good point to show the entire shell script and go through the details at the end of the listing. The SSAidentify.ksh shell script is shown in Listing 20.9.
#!/bin/ksh # # SCRIPT: SSAidentify.ksh # # AUTHOR: Randy Michael # # DATE: 11/7/2000

Listing 20.9 SSA identify.ksh shell script listing. (continues)

Free & Share & Open

508

Chapter 20

# # # # # # # # # # # # # # # # # # # # # # # # # #

REV: 2.5.A PURPOSE: This script is used to turn on, or off, the identify lights on the system’s SSA disks REV LIST: 11/27/2000: Added code to allow the user to turn on/off individual pdisk lights 12/10/2000: Added code to accept a combination of pdisks and hdisks. For each hdisk passed the script translates the hdisk# into the associated pdisk#(s). 12/10/2000: Added code to ALLOW using the currently VARIED ON Volume Group’s disks (-v switch), as opposed to ALL DEFINED SSA disks, which is the default behavior. Very helpful in an HACMP environment. 12/11/2000: Added the “twirl” function to give the user feedback during long processing periods, i.e., translating a few hundred hdisks into associated pdisks. The twirl function is just a rotating cursor, and it twirls during the translation processing. set -n set -x # Uncomment to check syntax without any execution # Uncomment to debug this script

SCRIPTNAME=$(basename $0) ############################################## function usage { echo “\nUSAGE ERROR... \nMAN PAGE ==> $SCRIPTNAME -? \nTo Turn ALL Lights Either ON or OFF: \nUSAGE: SSAidentify.ksh [-v] [on] [off] EXAMPLE: SSAidentify.ksh -v on \nWill turn ON ALL of the system’s currently VARIED ON SSA identify lights. NOTE: The default is all DEFINED SSA disks \nTo Turn SPECIFIC LIGHTS Either ON or OFF Using EITHER the pdisk#(s) AND/OR the hdisk#(s): \nUSAGE: SSAidentify.ksh [on] [off] pdisk{#1} [hdisk{#2}]... EXAMPLE: SSAidentify.ksh on hdisk36 pdisk44 pdisk47 \nWill turn ON the lights to all of the associated pdisk#(s) that hdisk36 translates to and PDISKS pdisk44 and pdisk47. \nNOTE: Can use all pdisks, all hdisks or BOTH hdisk

Listing 20.9 SSA identify.ksh shell script listing. (continued)

Turning On/Off SSA Identification Lights

509

and pdisk together if you want...” exit 1 } ############################################## function man_page { MAN_FILE=”/tmp/man_file.out” >$MAN_FILE # Text for the man page... echo “\n\t\tMAN PAGE FOR SSAidentify.ksh SHELL SCRIPT\n This script is used to turn on, or off, the system’s SSA disk drive identification lights. You can use this script in the following ways:\n To turn on/off ALL DEFINED SSA drive identification lights, ALL VARIEDON SSA drive identification lights (-v switch), AN INDIVIDUAL SSA drive identification light or A LIST OF SSA drive identification lights.\n SSA disk drives can be specified by EITHER the pdisk OR the hdisk, or a COMBINATION OF BOTH. The script translates all hdisks into the associated pdisk(s) using the system’s /usr/sbin/ssaxlate command and turns the SSA identification light on/off using the system’s /usr/sbin/ssaidentify command.\n This script has four switches that control its action:\n -? - Displays this man page.\n on - Turns the SSA identify light(s) ON.\n off - Turns the SSA identify light(s) OFF.\n -v - Specifies to only act on SSA disks that are in currently varied-on volume groups. The default action is to act on ALL DEFINED SSA disks.\n NOTE: This switch is ignored for turning on/off individual SSA drive lights, only valid when turning on/off ALL lights. This option is very helpful in an HACMP environment because ALL DEFINED, the default action, will turn on/off all of the SSA drive lights even if the SSA disk is in a volume group that is not currently varied-on. This can be confusing in an HA cluster.\n Using this script is very straight forward. The following examples show the

Listing 20.9 SSA identify.ksh shell script listing. (continues)

Free & Share & Open

510

Chapter 20

correct use of this script:\n” >> $MAN_FILE echo “\nUSAGE: SSAidentify.ksh [-v] [on] [off] [pdisk#/hdisk#] [pdisk#/hdisk# list] \n\nTo Turn ALL Lights Either ON or OFF: \nUSAGE: SSAidentify.ksh [-v] [on] [off] \nEXAMPLE: $SCRIPTNAME on \nWill turn ON ALL of the system’s DEFINED SSA identify lights. This is the default. EXAMPLE: SSAidentify.ksh -v on \nWill turn ON ALL of the system’s currently VARIED-ON SSA identify lights. OVERRIDES THE DEFAULT ACTION OF ALL DEFINED SSA DISKS \nTo Turn SPECIFIC LIGHTS Either ON or OFF Using EITHER the pdisk#(s) AND/OR the hdisk#(s): \nUSAGE: $SCRIPTNAME [on] [off] pdisk{#1} [hdisk{#2}]... \nEXAMPLE: $SCRIPTNAME on hdisk36 pdisk44 pdisk47 \nWill turn ON the lights to all of the associated pdisk#(s) that hdisk36 translates to and PDISKS pdisk44 and pdisk47. \nNOTE: Can use all pdisks, all hdisks or BOTH hdisk and pdisk together if you want...\n\n” >> $MAN_FILE more $MAN_FILE # End of man_page function } ############################################## function cleanup { echo “\n...Exiting on a trapped signal...EXITING STAGE LEFT...\n” kill $TWIRL_PID # End of cleanup function } ############################################## function twirl { TCOUNT=”0”

# For each TCOUNT the line twirls one increment

while : # Loop forever...until you break out of the loop do TCOUNT=$(expr ${TCOUNT} + 1) # Increment the TCOUNT case ${TCOUNT} in

Listing 20.9 SSA identify.ksh shell script listing. (continued)

Turning On/Off SSA Identification Lights

511

“1”)

“2”)

“3”)

“4”)

*) esac

echo ‘-’”\b\c” sleep 1 ;; echo ‘\\’”\b\c” sleep 1 ;; echo “|\b\c” sleep 1 ;; echo “/\b\c” sleep 1 ;; TCOUNT=”0” ;; # Reset the TCOUNT to “0”, zero.

done # End of twirl function } ############################################ function all_defined_pdisks { # TURN ON/OFF ALL LIGHTS: # Loop through each of the system’s pdisks by using the “lsdev” # command with the “-Cc pdisk” switch while using “awk” to extract # out the actual pdisk number. We will either # turn the identifier lights on or off specified by the # $SWITCH variable: # # Turn lights on: -y # Turn lights off: -n # # as the $SWITCH value to the “ssaidentify” command, as used below... echo “\nTurning $STATE ALL of the system’s pdisks...Please Wait...\n” for PDISK in $(lsdev -Cc pdisk -s ssar -H | awk ‘{print $1}’ | grep pdisk) do echo “Turning $STATE ==> $PDISK” ssaidentify -l $PDISK -${SWITCH} || echo “Turning $STATE $PDISK Failed” done echo “\n...TASK COMPLETE...\n” } ############################################

Listing 20.9 SSA identify.ksh shell script listing. (continues)

Free & Share & Open

512

Chapter 20

function all_varied_on_pdisks { trap ‘kill -9 $TWIRL_PID; return 1’ 1 2 3 15 cat /dev/null > $HDISKFILE echo “\nGathering a list of Varied on system SSA disks...Please wait...\c” VG_LIST=$(lsvg -o) # Get the list of Varied ON Volume Groups for VG in $(echo $VG_LIST) do lspv | grep $VG >> $HDISKFILE # List of Varied ON PVs done twirl & # Gives the user some feedback during long processing times... TWIRL_PID=$! echo “\nTranslating hdisk(s) into the associated pdisk(s)...Please Wait... \c” for DISK in $(cat $HDISKFILE) # Translate hdisk# into pdisk#(s) do # Checking for an SSA disk /usr/sbin/ssaxlate -l $DISK # 2>/dev/null 1>/dev/null if (($? == 0)) then /usr/sbin/ssaxlate -l $DISK >> $PDISKFILE # Add to pdisk List fi done kill -9 $TWIRL_PID # Kill the user feedback function... echo “\b “

echo “\nTurning $STATE all VARIED-ON system pdisks...Please Wait...\n” for PDISK in $(cat $PDISKFILE) do # Act on each pdisk individually... echo “Turning $STATE ==> $PDISK” /usr/sbin/ssaidentify -l $PDISK -${SWITCH} || echo “Turning $STATE $PDISK Failed”

Listing 20.9 SSA identify.ksh shell script listing. (continued)

Turning On/Off SSA Identification Lights

513

done echo “\n\t...TASK COMPLETE...\n” } ############################################ function list_of_disks { # TURN ON/OFF INDIVDUAL LIGHTS: # Loop through each of the disks that was passed to this script # via the positional parameters greater than $1, i.e., $2, $3, $4... # We first determine if each of the parameters is a pdisk or an hdisk. # For each hdisk passed to the script we first need to translate # the hdisk definition into a pdisk definition. This script has # been set up to accept a combination of hdisks and pdisks. # # We will either turn the identifier lights on or off specified by # the $SWITCH variable for each pdisk#: # # Turn lights on: -y # Turn lights off: -n # # as the $SWITCH value to the “ssaidentify” command. echo “\n” # The disks passed to this script can be all hdisks, all pdisks # or a combination of pdisks and hdisks; it just does not matter. # We translate each hdisk into the associated pdisk(s).

echo “\nTurning $STATE individual SSA disk lights...\n” for PDISK do # Is if [ then in $(echo $PDISKLIST) it a real pdisk?? -c /dev/${PDISK} ] 2>/dev/null # Yep - act on it... /usr/sbin/ssaidentify -l $PDISK -${SWITCH} if [ $? -eq 0 ] then /usr/bin/ssaxlate -l $PDISK -${SWITCH} if (($? == 0))

Listing 20.9 SSA identify.ksh shell script listing. (continues)

Free & Share & Open

514

Chapter 20

then echo “Light on $PDISK is $STATE” else echo “Turning $STATE $PDISK Failed” fi fi else echo “\nERROR: $PDISK is not a defined device on $THISHOST\n” fi done echo “\n...TASK COMPLETE...\n” } ############################################ ############# BEGINNING OF MAIN ############ ############################################ # Set a trap... # Remember...Cannot trap a “kill -9” !!! trap ‘cleanup;exit 1’ 1 2 3 15 ############################################## # Check for the correct number of arguments (1) if (($# == 0)) then usage fi ############################################## # See if the system has any pdisks defined before proceeding PCOUNT=$(lsdev -Cc pdisk -s ssar | grep -c pdisk) if ((PCOUNT == 0)) then echo “\nERROR: This system has no SSA disks defined\n” echo “\t\t...EXITING...\n”

Listing 20.9 SSA identify.ksh shell script listing. (continued)

Turning On/Off SSA Identification Lights

515

exit 1 fi ############################################## # Make sure that the ssaidentify program is # executable on this system... if [ ! -x then echo echo echo exit fi /usr/sbin/ssaidentify ] “\nERROR: /usr/sbin/ssaidentify is NOT an executable” “program on $THISHOST” “\n...EXITING...\n” 1

############################################## # Make sure that the ssaxlate program is # executable on this system... if [ ! -x then echo echo echo exit fi /usr/sbin/ssaxlate ] “\nERROR: /usr/sbin/ssaxlate is NOT an executable” “program on $THISHOST” “\n...EXITING...\n” 1

############################################## ############################################## # # Okay, we should have valid data at this point # Let’s do a light show. # ############################################## ############################################## # Always use the UPPERCASED value for the $STATE, $MODE, # and $PASSED variables...

typeset -u MODE

Listing 20.9 SSA identify.ksh shell script listing. (continues)

Free & Share & Open

516

Chapter 20

MODE=”DEFINED_DISKS” typeset -u STATE STATE=UNKNOWN typeset -u PASSED # Use lowercase for the argument list typeset -l ARGUMENT # Grab the system hostname THISHOST=$(hostname) # Define the hdisk and pdisk FILES HDISKFILE=”/tmp/disklist.out” >$HDISKFILE PDISKFILE=”/tmp/pdisklist.identify” >$PDISKFILE # Define the hdisk and pdisk list VARIABLES HDISKLIST= PDISKLIST= # Use getopts to parse the command-line arguments while getopts “:v V” ARGUMENT 2>/dev/null do case $ARGUMENT in v|V) \?) esac done ############################################## # Decide if we are to turn the lights on or off... (echo $@ | grep -i -w on >/dev/null) && STATE=ON (echo $@ | grep -i -w off >/dev/null) && STATE=OFF MODE=”VARIED_ON” ;; man_page ;;

case $STATE in ON)

Listing 20.9 SSA identify.ksh shell script listing. (continued)

Turning On/Off SSA Identification Lights

517

# Turn all of the lights ON... SWITCH=”y” ;; OFF) # Turn all of the lights OFF... SWITCH=”n” ;; *) # Unknown Option... echo “\nERROR: Please indicate the action to turn lights ON or OFF\n” usage exit 1 ;; esac ############################################## ############################################## ########## PLAY WITH THE LIGHTS ############## ############################################## ############################################## if (($# == 1)) && [[ $MODE = “DEFINED_DISKS” ]] then # This function will turn all lights on/off all_defined_pdisks elif [[ $MODE = “VARIED_ON” ]] && (($# = 2)) then # This function will turn on/off SSA disk lights # in currently varied-on volume groups only all_varied_on_pdisks # Now check for hdisk and pdisk arguments elif [ $MODE = DEFINED_DISKS ] && (echo $@ | grep disk >/dev/null) \ && (($# >= 2)) then # If we are here we must have a list of hdisks # and/or pdisks # Look for hdisks and pdisks in the command-line arguments for DISK in $(echo $@ | grep disk) do case $DISK in

Listing 20.9 SSA identify.ksh shell script listing. (continues)

Free & Share & Open

518

Chapter 20

hdisk*) HDISKLIST=”$HDISKLIST $DISK” ;; pdisk*) PDISKLIST=”$PDISKLIST $DISK” ;; *) : # No-Op - Do nothing ;; esac done if [[ ! -z “$HDISKLIST” ]] # Check for hdisks to convert to pdisks then # We have some hdisks that need to be converted to pdisks # so start converting the hdisks to pdisks # Give the user some feedback echo “\nConverting hdisks to pdisk definitions” echo “\n ...Please be patient...\n” # Start converting the hdisks to pdisks for HDISK in $(echo $HDISKLIST) do PDISK=$(ssaxlate -l $HDISK) if (($? == 0)) then echo “$HDISK translates to ${PDISK}” else echo “ERROR: hdisk to pdisk translation FAILED for $HDISK” fi # Build a list of pdisks # Add pdisk to the pdisk list PDISKLIST=”$PDISKLIST $PDISK” done fi if [[ -z “$PDISKLIST” ]] then echo “\nERROR: You must specify at least one hdisk or pdisk\n” man_page exit 1

Listing 20.9 SSA identify.ksh shell script listing. (continued)

Turning On/Off SSA Identification Lights

519

else # Turn on/off the SSA identification lights list_of_disks fi fi ############################################## # END OF SCRIPT # ##############################################

Listing 20.9 SSA identify.ksh shell script listing. (continued)

Let’s start at the “Beginning of Main” in Listing 20.9. The very first thing that we do is set a trap. This trap is set for exit codes 1, 2, 3, 5, and 15. On any of these captured signals the cleanup function is executed, and then the shell script exits with a return code of 1. It is nice to be able to clean up before the shell script just exits. In the next series of tests we first make sure that there is at least one argument present on the command line. If no arguments are given, then the script presents the usage function, which displays proper usage and exits. If we pass the argument test then I thought it would be a good idea to see if the system has any SSA disks defined on the system. For this step we use the PCOUNT=$(lsdev -Cc pdisk -s ssar | grep -c pdisk). The grep -c returns the count of SSA disks found on the system and assigns the value to the PCOUNT variable. If the value is zero there are no SSA disks, so inform the user and exit. If we do have some SSA disks, the next thing we do is make sure that the ssaidentify and ssaxlate commands exist and are executable on this system. At this point we know we are in an SSA environment so we define and initialize all of the script’s variables. Then we get to use the getopts function to parse the command-line arguments. We expect and recognize just two arguments, -v and -V, to specify varied-on volume groups only. Any other argument, specified by a preceding hyphen, -, displays the man_page function. Anything else on the command line is ignored by the getopts function, which is a shell built-in function. On the command line we must have either on or off present, or we do not have enough information to do anything. We check the command-line arguments by echoing out the full list and grepping for on and off. At the next case statement the $STATE variable is tested. If on or off was not found, the usage function is displayed and the script exits. If we get past this point we know that we have the minimal data to do some work. When we start playing with the lights we have to do some tests to decide what action we need to take and on what set of SSA disks. The first one is simple. If we have only one command-line argument and it is either on or off, then we know to turn on or

Free & Share & Open

520

Chapter 20 off all defined SSA disk identification lights on the system without regard to volume group status. So, here we run the all_defined_pdisks function. If we have two arguments on the command line and one of them is -v or -V, then we know to act only on SSA disks in currently varied-on volume groups by turning every one of the SSA identification lights on or off. The last option is to have hdisks or pdisks listed on the command line. For this option we know to act on only the disks that the user specified and to turn on or off only these disks. Because we allow both hdisks and pdisks we need to convert everything to pdisk definitions before we call the list_of_disks function. To do this we echo the entire list of command-line arguments and grep for the word disk. Using this list in a case statement, we can detect the presence of an hdisk or a pdisk. For each one found it is added to either the HDISKLIST or the PDISKLIST variables. After the test we check to see if the HDISKLIST variable has anything assigned, which means that the variable is not null. If there are entries, then we convert each hdisk to its associated pdisk and build up the pdisk list in the PDISKLIST variable. When the list is complete, and it is not an empty list, we call the list_of_disks function. That is it for this shell script.

Other Things to Consider
I cannot always fit all of the options into a chapter, and this chapter is no exception. Here are a few things to consider to modify this shell script.

Error Log
When I created this shell script it was for a personal need because I have so many SSA disk trays. For my purposes I did not need an error log, but you may find one necessary. In the places that I sent everything to the bit bucket, especially standard error, or file descriptor 2, redirect this output to append to an error log. This may help you find something in the system that you missed.

Cross-Reference
Because it is rare for the hdisk and pdisk associations to match by numbers you may find that a shell script to cross-reference the numbers beneficial. You should be able to knock this out in about one hour. Look through the code where I first test the hdisk to see if it is an SSA disk and then do the translation. Using these few lines of code you can build a nice little cross-reference sheet for your staff.

Root Access and sudo
Both of the SSA commands need root privilege to execute. If your systems have strict root access rules you may just want to define this shell script in your /etc/sudoers file. Please never directly edit this file! There is a special wrapper program around the

Turning On/Off SSA Identification Lights vi editor in the /usr/local/sbin directory called visudo. This command starts a vi session and opens up the /etc/sudoers file automatically. When you are finished editing and save the file, this program checks the /etc/sudoers file for errors.

521

Summary
In this chapter we learned a few new things about controlling the SSA subsystem on an AIX machine. These principles apply to any other Unix system utilizing SSA. As always, there are many different ways to write a shell script, and some are lean and mean with no comments. I like to make the shell scripts easier to understand and maintain. But I do have a few things that you may want to consider. I hope you learned something in this chapter. In the next chapter we will look at pseudo-random number generators. See you in the next chapter!

Free & Share & Open

CHAPTER

21
Pseudo-Random Number Generation

In writing shell scripts we sometimes run into a situation where we are creating files faster than we can make the filenames unique. Most of the time a date/time stamp can be added as a suffix to the filename to make the filename unique, but if we are creating more than one file per second we end up overwriting the same file during a single second. To get around this problem we can create pseudo-random numbers to append to the filename after the date/time stamp. You may recall that in Chapter 10 we studied creating pseudo-random passwords by using the computer generated numbers as pointers to array elements that contained keyboard characters. A more thorough discussion of randomness is presented in this chapter.

What Makes a Random Number?
It is very difficult to create a true random number in a computer system. The problem is repeatability and predictability of the number. When you start researching random numbers you quickly enter the realm of heavy mathematical theory, and many of the researchers have varying opinions of randomness. The only true random numbers that I know of are the frequency variations of radioactive decay events and the frequency variations of white noise. Radioactive decay events would have to be detected in some way, and because we do not want to have any radioactive material hanging around we can use built-in computer programs called pseudo-random number generators. Some

523

Free & Share & Open

524

Chapter 21 computer techniques are able to create numbers that are suitable for encryption keys and for cryptographic secure communication links. Some of these techniques include measuring the time between keystrokes and use this measured value as a memory address to read the contents. A popular Unix technique is to use a special Unix character device called /dev/ random. If you search the Internet for /dev/random you will find more information than you could imagine on the topic of randomness. Randomness is a discussion topic with many experts in the field having widely varying viewpoints. I am not an expert on randomness, and this topic is beyond the scope of this book. We are going to concentrate on creating pseudo-random numbers to make unique filenames in this chapter. The numbers that we will create are not sufficiently random for any type of encryption because they are repeatable and cyclical in nature, but they will create unique filenames. The Korn shell provides an environment variable called—you guessed it— RANDOM. This pseudo-random number generator uses a seed as a starting point to create all future numbers in the sequence. After the initial seed is used to create a pseudo-random number, this resulting number is used for the next seed to create the next random number, and so on. As you would expect, if you always start generating your numbers with the same seed each time, you will get the exact same number sequence each time. To change the repeatability we need to have a mechanism to vary the initial seed each time we start generating numbers. I like to use the current process ID (PID) because this number will vary widely and is an easy way to change the seed value each time we start generating numbers.

The Methods
In this chapter we are going to look at three techniques to generate pseudo-random numbers:
■ ■

Create numbers between zero and the maximum number allowed by the system (32,767) Create numbers between one and a user-defined maximum Create fixed-length numbers between one and a user-defined maximum with leading zeros added if needed

■ ■ ■ ■

Each method is valid for a filename extension, but you may have other uses that require either a range of numbers or a fixed number of digits with leading zeros. In any case the basic concept is the same. We start out by initializing the RANDOM environment variable to the current PID:
RANDOM=$$

The double dollar signs ($$) specify the PID of the current system process. The PID will vary, so this is a good way to initialize the RANDOM variable so that we do not always repeat the same number sequence. Once the RANDOM environment variable is initialized we can use RANDOM just like any other variable. Most of the time we will use

Pseudo-Random Number Generation the echo command to print the next pseudo-random number. An example of using the RANDOM environment variable is shown in Listing 21.1.
# RANDOM=$$ # echo $RANDOM 23775 # echo $RANDOM 3431 # echo $RANDOM 12127 # echo $RANDOM $RANDOM 2087 21108

525

Listing 21.1 Using the RANDOM environment variable.

By default the RANDOM variable will produce numbers between 0 and 32767. Notice the last entry in Listing 21.1. We can produce more than one number at a time if we need to by adding more $RANDOM entries to our echo command. In showing our three methods in this chapter we are going to create three functions and then create a shell script that will use one of the three methods depending on the user-supplied input. The last step is to write a shell script that will create unique filenames using a date/ time stamp and a random number.

Method 1: Creating Numbers between 0 and 32,767
Creating pseudo-random numbers using this default method is the simplest way to use the RANDOM environment variable. The only thing that we need to do is to initialize the RANDOM environment variable to an initial seed value and use the echo command to display the new number. The numbers will range from 0 to 32767, which is the maximum for the RANDOM variable. You do not have control over the number of digits, except that the number of digits will not exceed five, and you cannot specify a maximum value for the number in this first method. The function get_random_number is shown in Listing 21.2. function get_random_number { # This function gets the next random number from the # $RANDOM variable. The range is 0 to 32767. echo “$RANDOM” }

Listing 21.2 get_random_number function listing.

Free & Share & Open

526

Chapter 21
As you can see, the function is just one line, and we are assuming that the RANDOM environment variable is initialized in the main body of the calling shell script.

Method 2: Creating Numbers between 1 and a User-Defined Maximum
We often want to limit the range of numbers to not exceed a user-defined maximum. An example is creating lottery numbers between 1 and the maximum number, which might be 36. We are going to use the modulo arithmetic operator to reduce all numbers to a fixed set of numbers between [0..N-1], which is called modulo N arithmetic. For our number range we need a user-supplied maximum value, which we will assign to a variable called UPPER_LIMIT. The modulo operator is the percent sign (%), and we use this operator the same way that you use the forward slash (/) in division. We still use the RANDOM environment variable to get a new pseudo-random number. This time, though, we are going to use the following equation to limit the number to not exceed the user-defined maximum modulo N arithmetic.
RANDOM_NUMBER=$(($RANDOM % $UPPER_LIMIT + 1))

Notice that we added one to the equation. Using the preceding equation will produce a pseudo-random number between 1 and the user-defined $UPPER_LIMIT. The function using this equation is in_range_random_number and is shown in Listing 21.3. function in_range_random_number { # Create a pseudo-random number less than or equal # to the $UPPER_LIMIT value, which is user defined RANDOM_NUMBER=$(($RANDOM % $UPPER_LIMIT + 1)) echo “$RANDOM_NUMBER” }

Listing 21.3 in_range_random_number function listing.

The function in Listing 21.3 assumes that the RANDOM variable seed has been initialized in the main body of the calling shell script and that a user-defined UPPER_LIMIT variable has been set. This function will produce numbers between 1 and the userdefined maximum value, but the number of digits will vary as the numbers are produced.

Pseudo-Random Number Generation

527

Method 3: Fixed-Length Numbers between 1 and a User-Defined Maximum
In both of the previous two examples we had no way of knowing how many digits the new number would contain. When we are creating unique filenames it would be nice to have filenames that are consistent in length. We can produce fixed-length numbers by padding the number with leading zeros for each missing digit. As an example we want all of our numbers to have four digits. Now let’s assume that the number that is produced is 24. Because we want 24 to have four digits, we need to pad the number with two leading zeros, which will make the number 0024. To pad the number we need to know the length of the character string that makes up the number. The Korn shell uses the pound operator (#) preceding the variable enclosed within curly braces ({}), as shown here.
RN_LENGTH=$(echo ${#RANDOM_NUMBER})

If the RANDOM_NUMBER variable has 24 assigned as an assigned value, then the result of the previous command is 2 (this RN_LENGTH variable points to the value 2), indicating two digits. We will also need the length of the UPPER_LIMIT value, and we will use the difference to know how many zeros to use to pad the pseudo-random number output. Take a close look at the code in Listing 21.4 where you will find the function in_range_fixed_length_random_number. function in_range_fixed_length_random_number { # Create a pseudo-random number less than or equal # to the $UPPER_LIMIT value, which is user defined. # This function will also pad the output with leading # zeros to keep the number of digits consistent. RANDOM_NUMBER=$(($RANDOM % $UPPER_LIMIT + 1)) # Find the length of each character string RN_LENGTH=$(echo ${#RANDOM_NUMBER}) UL_LENGTH=$(echo ${#UPPER_LIMIT}) # Calculate the difference in string length (( LENGTH_DIFF = UL_LENGTH - RN_LENGTH )) # Pad the $RANDOM_NUMBER value with leading zeros

Listing 21.4 in_range_fixed_length_random_number function. (continues)

Free & Share & Open

528

Chapter 21

# to keep the number of digits consistent. case $LENGTH_DIFF in 0) echo “$RANDOM_NUMBER” ;; 1) echo “0$RANDOM_NUMBER” ;; 2) echo “00$RANDOM_NUMBER” ;; 3) echo “000$RANDOM_NUMBER” ;; 4) echo “0000$RANDOM_NUMBER” ;; 5) echo “00000$RANDOM_NUMBER” ;; *) echo “$RANDOM_NUMBER” ;; esac }

Listing 21.4 in_range_fixed_length_random_number function. (continued)

In Listing 21.4 we use the same technique from Listing 21.3 to set an upper limit to create our numbers, but we add in code to find the string length of both the UPPER_LIMIT and RANDOM_NUMBER values. By knowing the length of both strings we subtract the random-number length from the upper-limit length and use the difference in a case statement to add the correct number of zeros to the output. Because this is a function, we again need to assume that the UPPER_LIMIT is defined and the RANDOM environment variable is initialized in the main body of the calling shell script. The resulting output is a fixed-length pseudo-random number padded with leading zeros if the output string length is less than the upper limit string length. Example output is shown in Listing 21.5 for an UPPER_LIMIT value of 9999.
0024 3145 9301 0328 0004 4029 2011 0295 0159 4863

Listing 21.5 Sample output for fixed-length random numbers.

Pseudo-Random Number Generation

529

Why Pad the Number with Zeros the Hard Way?
An easier, and much cleaner, way to pad a number with leading zeros is to typeset the variable to a fixed length. The following command works for any length number: typeset -Z5 FIXED_LENGTH FIXED_LENGTH=25 echo $FIXED_LENGTH 00025

Listing 21.6 Using the typeset command to fix the length of a variable.

In the example in Listing 21.6 we used the typeset command to set the length of the FIXED_LENGTH variable to five digits. Then we assigned the value 25 to it. When we use the echo command to show the value assigned to the variable the result is 00025, which is fixed to five digits. Let’s modify the function in Listing 21.4 to use this technique as shown in Listing 21.7. function in_range_fixed_length_random_number_typeset { # Create a pseudo-random number less than or equal # to the $UPPER_LIMIT value, which is user defined. # This function will also pad the output with leading # zeros to keep the number of digits consistent using # the typeset command. # Find the length of each character string UL_LENGTH=$(echo ${#UPPER_LIMIT}) # Fix the length of the RANDOM_NUMBER variable to # the length of the UPPER_LIMIT variable, specified # by the $UL_LENGTH variable. typeset -Z$UL_LENGTH RANDOM_NUMBER

# Create a fixed length pseudo-random number RANDOM_NUMBER=$(($RANDOM % $UPPER_LIMIT + 1)) # Return the value of the fixed length $RANDOM_NUMBER echo $RANDOM_NUMBER }

Listing 21.7 Using the typeset command in a random number function.

Free & Share & Open

530

Chapter 21
As you can see in Listing 21.7, we took all of the complexity out of fixing the length of a number. The only value we need to know is the length of the UPPER_LIMIT variable assignment. As an example, if the upper limit is 9999 then the length is 4. We use 4 to typeset the RANDOM_NUMBER variable to four digits. Now that we have four functions that will create pseudo-random numbers, we can proceed with a shell script that will use one of the three methods depending on the command-line arguments supplied to the shell script.

Shell Script to Create Pseudo-Random Numbers
Using the three functions from Listings 21.2, 21.3, and 21.7 we are going to create a shell script that, depending on the command-line arguments, will use one of these three functions. We first need to define how we are going to use each function. With the usage definitions from Table 21.1 let’s create a shell script. We already have the functions to create the numbers so we will start with BEGINNING OF MAIN in the shell script. For the usage function we will need the name of the shell script. We never want to hard-code the name of a shell script because someone may rename the shell script for one reason or another. To query the system for the actual name of the shell script we use the basename $0 command. This command will return the name of the shell script, specified by the $0 argument, with the directory path stripped out. I like to use either of the following commands to create a SCRIPT_NAME variable.

SCRIPT_NAME=`basename $0`

or
SCRIPT_NAME=$(basename $0)

Table 21.1

random_number.ksh Shell Script Usage FUNCTION USED TO CREATE THE NUMBER Without argument will use get_random_number With one numeric argument will use in_range_random_number With -f as the first argument followed by a numeric argument will use in_range_fixed_length_random_number_typeset

SHELL SCRIPT USAGE random_number.ksh random_number.ksh 9999 random_number.ksh -f 9999

Pseudo-Random Number Generation
The result of both command substitution commands is the same. Next we need to initialize the RANDOM environment variable. As we described before, we are going to use the current process ID as the initial seed for the RANDOM variable.
RANDOM=$$

531

The SCRIPT_NAME and the RANDOM variables are the only initialization needed for this shell script. The rest of the script is a case statement that uses the number of command-line arguments ($#) as a value to decide which random number function we will use. We also do some numeric tests to ensure that “numbers” are actually numeric values. For the numeric tests we use the regular expression +([0-9]) in a case statement. If the value is a number, then we do nothing, which is specified by the no-op character, colon (:). The entire shell script is shown in Listing 21.8.
#!/usr/bin/ksh # # AUTHOR: Randy Michael # SCRIPT: random_number.ksh # DATE: 11/12/2001 # REV: 1.2.P # # PLATFORM: Not Platform Dependent # # EXIT CODES: # 0 - Normal script execution # 1 - Usage error # # REV LIST: # # # set -x # Uncomment to debug # set -n # Uncomment to check syntax without any command execution # #################################################### ########## DEFINE FUNCTIONS HERE ################### #################################################### function usage { echo “\nUSAGE: $SCRIPT_NAME [-f] [upper_number_range]” echo “\nEXAMPLE: $SCRIPT_NAME” echo “Will return a random number between 0 and 32767” echo “\nEXAMPLE: $SCRIPT_NAME 1000” echo “Will return a random number between 1 and 1000” echo “\nEXAMPLE: $SCRIPT_NAME -f 1000”

Listing 21.8 random_number.ksh shell script listing. (continues)

Free & Share & Open

532

Chapter 21

echo “Will add leading zeros to a random number from” echo “1 to 1000, which keeps the number of digits consistant\n” } #################################################### function get_random_number { # This function gets the next random number from the # $RANDOM variable. The range is 0 to 32767. echo “$RANDOM” } #################################################### function in_range_random_number { # Create a pseudo-random number less than or equal # to the $UPPER_LIMIT value, which is user defined RANDOM_NUMBER=$(($RANDOM % $UPPER_LIMIT + 1)) echo “$RANDOM_NUMBER” } ####################################################

function in_range_fixed_length_random_number_typeset { # Create a pseudo-random number less than or equal # to the $UPPER_LIMIT value, which is user defined. # This function will also pad the output with leading # zeros to keep the number of digits consistent using # the typeset command. # Find the length of each character string UL_LENGTH=$(echo ${#UPPER_LIMIT}) # Fix the length of the RANDOM_NUMBER variable to # the length of the UPPER_LIMIT variable, specified

Listing 21.8 random_number.ksh shell script listing. (continued)

Pseudo-Random Number Generation

533

# by the $UL_LENGTH variable. typeset -Z$UL_LENGTH RANDOM_NUMBER

# Create a fixed length pseudo-random number RANDOM_NUMBER=$(($RANDOM % $UPPER_LIMIT + 1)) # Return the value of the fixed length $RANDOM_NUMBER echo $RANDOM_NUMBER }

#################################################### ############## BEGINNING OF MAIN ################### #################################################### SCRIPT_NAME=`basename $0` RANDOM=$$ # Initialize the RANDOM environment variable # using the PID as the initial seed

case $# in 0) get_random_number ;; 1) UPPER_LIMIT=”$1” # Test to see if $UPPER_LIMIT is a number case $UPPER_LIMIT in +([0-9])) : # Do Nothing...It’s a number # NOTE: A colon (:) is a no-op in Korn shell ;; *) echo “\nERROR: $UPPER_LIMIT is not a number...” usage exit 1 ;; esac # We have a valid UPPER_LIMIT. Get the number. in_range_random_number

Listing 21.8 random_number.ksh shell script listing. (continues)

Free & Share & Open

534

Chapter 21

;; 2) # Check for the -f switch to fix the length. if [[ $1 = ‘-f’ ]] || [[ $1 = ‘-F’ ]] then UPPER_LIMIT=”$2” # Test to see if $UPPER_LIMIT is a number case $UPPER_LIMIT in +([0-9])) : # Do Nothing...It’s a number # NOTE: A colon (:) is a no-op in Korn shell ;; *) echo “\nERROR: $UPPER_LIMIT is not a number...” usage exit 1 ;; esac in_range_fixed_length_random_number_typeset else echo “\nInvalid argument $1, see usage below...” usage exit 1 fi ;; *) usage exit 1

;; esac # End of random_number.ksh Shell Script

Listing 21.8 random_number.ksh shell script listing. (continued)

Notice in Listing 21.8 that we will allow only zero, one, or two command-line arguments. More than three arguments produces an error, and nonnumeric values, other than -f or -F in argument one, will produce a usage error. Output using the random_number.ksh shell script is shown in Listing 21.9.

Pseudo-Random Number Generation

535

yogi@/scripts# 10859 yogi@/scripts# 14493 yogi@/scripts# 05402 yogi@/scripts#

random_number.ksh 32000 random_number.ksh -f 32000 ./random_number.ksh -f 32000 ./random_number.ksh -f

ERROR: -f is not a number... USAGE: random_number.ksh [-f] [upper_number_range] EXAMPLE: random_number.ksh Will return a random number between 0 and 32767 EXAMPLE: random_number.ksh 1000 Will return a random number between 1 and 1000 EXAMPLE: random_number.ksh -f 1000 Will add leading zeros to a random number from 1 to 1000, which keeps the number of digits consistent

Listing 21.9 random_number.ksh shell script in action.

The last part of Listing 21.9 is a usage error. Notice that we give an example of each of the three uses for the random_number.ksh shell script as well as state why the usage error occurred. Now that we have the shell script to produce pseudo-random numbers, we need to move on to creating unique filenames.

Creating Unique Filenames
The goal of this chapter is to write a shell script that will produce unique filenames using a date/time stamp with a pseudo-random number as an extended suffix. When I create these unique filenames I like to keep the length of the filenames consistent so we are going to use only one of the random number functions, in_range_fixed_ length_random_number_typeset. We have a few new pieces to put into this new shell script. First we have to assume that there is some program or shell script that will be putting data into each of the unique files. To take care of executing the program or shell script we can add a function

Free & Share & Open

536

Chapter 21 that will call the external program, and we will redirect our output to the new unique filename on each loop iteration. The second piece is that we need to ensure that we never use the same number during the same second. Otherwise, the filename is not unique and the data will be overwritten. We need to keep a list of each number that we use during each second and reset the USED_NUMBERS list to null on each new second. In addition we need to grep the list each time we create a new number to see if it has already been used. If the number has been used we just create a new number and check for previous usage again. The procedure to step through our new requirements is not difficult to understand once you look at the code. The full shell script is shown in Listing 21.10, and an example of using the shell script is shown in Listing 21.11. Please study the script carefully, and we will go through the details at the end.
#!/usr/bin/ksh # # AUTHOR: Randy Micahel # SCRIPT: mk_unique_filename.ksh # DATE: 11/12/2001 # REV: 1.2.P # # PLATFORM: Not Platform Dependent # # EXIT CODES: # 0 - Normal script execution # 1 - Usage error # # REV LIST: # # # set -x # Uncomment to debug # set -n # Uncomment to debug without any execution # #################################################### ########## DEFINE FUNCTIONS HERE ################### #################################################### function usage { echo “\nUSAGE: $SCRIPT_NAME base_file_name\n” exit 1 } #################################################### function get_date_time_stamp { DATE_STAMP=$(date +’%m%d%y.%H%M%S’)

Listing 21.10 mk_unique_filename.ksh shell script listing.

Pseudo-Random Number Generation

537

echo $DATE_STAMP } #################################################### function get_second { THIS_SECOND=$(date +%S) echo $THIS_SECOND } ####################################################

function in_range_fixed_length_random_number_typeset { # Create a pseudo-random number less than or equal # to the $UPPER_LIMIT value, which is user defined. # This function will also pad the output with leading # zeros to keep the number of digits consistent using # the typeset command. # Find the length of each character string UL_LENGTH=$(echo ${#UPPER_LIMIT}) # Fix the length of the RANDOM_NUMBER variable to # the length of the UPPER_LIMIT variable, specified # by the $UL_LENGTH variable. typeset -Z$UL_LENGTH RANDOM_NUMBER

# Create a fixed length pseudo-random number RANDOM_NUMBER=$(($RANDOM % $UPPER_LIMIT + 1)) # Return the value of the fixed length $RANDOM_NUMBER echo $RANDOM_NUMBER }

#################################################### function my_program {

Listing 21.10 mk_unique_filename.ksh shell script listing. (continues)

Free & Share & Open

538

Chapter 21

# Put anything you want to process in this function. I # recommend that you specify an external program of shell # script to execute. echo “HELLO WORLD - $DATE_ST” > $UNIQUE_FN & # } : # No-Op - Does nothing but has a return code of zero

#################################################### ################ BEGINNING OF MAIN ################# #################################################### SCRIPT_NAME=$(basename $0) # Query the system for this script name # Check for the correct number of arguments - exactly 1 if (( $# != 1 )) then echo “\nERROR: Usage error...EXITING...” usage fi # What filename do we need to make unique? BASE_FN=$1 RANDOM=$$ # Get the BASE filename to make unique # Initialize the RANDOM environment variable # with the current process ID (PID)

UPPER_LIMIT=32767 # Set the UPPER_LIMIT CURRENT_SECOND=99 # Initialize to a nonsecond LAST_SECOND=98 # Initialize to a nonsecond USED_NUMBERS= # Initialize to null

PROCESSING=”TRUE” # Initialize to run mode while [[ $PROCESSING = “TRUE” ]] do DATE_ST=$(get_date_time_stamp) # Get the current date/time CURRENT_SECOND=$(get_second) # Get the current second RN=$(in_range_fixed_length_random_number_typeset) # Get a new number # Check to see if we have already used this number this second if (( CURRENT_SECOND == LAST_SECOND ))

Listing 21.10 mk_unique_filename.ksh shell script listing. (continued)

Pseudo-Random Number Generation

539

then UNIQUE=FALSE # Initialize to FALSE while [[ “$UNIQUE” != “TRUE” ]] && [[ ! -z “$UNIQUE” ]] do # Has this number already been used this second? echo $USED_NUMBERS | grep $RN >/dev/null 2>&1 if (( $? == 0 )) then # Has been used...Get another number RN=$(in_range_fixed_length_random_number) else # Number is unique this second... UNIQUE=TRUE # Add this number to the used number list USED_NUMBERS=”$USED_NUMBERS $RN” fi done else USED_NUMBERS= # New second...Reinitialize to null fi # Assign the unique filename to the UNIQUE_FN variable UNIQUE_FN=${BASE_FN}.${DATE_ST}.$RN # echo $UNIQUE_FN # Comment out this line!! LAST_SECOND=$CURRENT_SECOND # Save the last second value # We have a unique filename... # # PROCESS SOMETHING HERE AND REDIRECT OUTPUT TO $UNIQUE_FN # my_program # # IF PROCESSING IS FINISHED ASSIGN “FALSE” to the # PROCESSING VARIABLE # # if [[ $MY_PROCESS = “done” ]] # then # PROCESSING=”FALSE” # fi done

Listing 21.10 mk_unique_filename.ksh shell script listing. (continued)

We need five functions in this shell script. As usual, we need a function for correct usage. We are expecting exactly one argument to this shell script, the base filename to make into a unique filename. The second function is used to get a date/time stamp.

Free & Share & Open

540

Chapter 21
The date command has a lot of command switches that allow for flexible date/time stamps. We are using two digits for month, day, year, hour, minute, and second with a period (.) between the date and time portions of the output. This structure is the first part that is appended to the base filename. The date command has the following syntax: date +/%m%d%y.%H%M%S'. We also need the current second of the current minute. The current second is used to ensure that the pseudo-random number that is created is unique to each second, thus a unique filename. The date command is used again using the following syntax: date +%S. The in_range_fixed_length_random_number_typeset function is used to create our pseudo-random numbers in this shell script. This function keeps the number of digits consistent for each number that is created. With the base filename, date/ time stamp, and the unique number put together, we are assured that every filename has the same number of characters. One more function is added to this shell script. The my_program function is used to point to the program or shell script that needs all of these unique filenames. It is better to point to an external program or shell script than trying to put everything in the internal my_program function and debugging the internal function on an already working shell script. Of course, I am making an assumption that you will execute the external program once during each loop iteration, which may not be the case. At any rate, this script will show the concept of creating unique filenames while remaining in a tight loop. At the BEGINNING OF MAIN in the main body of the shell script we first query the system for name of the shell script. The script name is needed for the usage function. Next we check for exactly one command-line argument. This single command-line argument is the base filename that we use to create further unique filenames. The next step is to assign our base filename to the variable BASE_FN for later use. The RANDOM environment variable is initialized with an initial seed, which we decided to be the current process ID (PID). This technique helps to ensure that the initial seed changes each time the shell script is executed. For this shell script we want to use the maximum value as the UPPER_LIMIT, which is 32767. If you need a longer or shorter pseudo-random number, you can change this value to anything you want. If you make this number longer than five digits the extra preceding digits will be zeros. There are four more variables that need to be initialized. We initialize both CURRENT_SECOND and LAST_SECOND to nonsecond values 99 and 98, respectively. The USED_NUMBERS list is initialized to null, and the PROCESSING variable is initialized to TRUE. The PROCESSING variable allows the loop to continue creating unique filenames and to keep calling the my_process function. Any non-TRUE value stops the loop and thus ends execution of the shell script. A while loop is next in our shell script, and this loop is where all of the work is done. We start out by getting a new date/time stamp and the current second on each loop iteration. Next a new pseudo-random number is created and is assigned to the RN variable. If the current second is the same as the last second, then we start another loop to ensure that the number that we created has not been previously used during the current second. It is highly unlikely that a duplicate number would be produced in such a short amount of time, but to be safe we need to do a sanity check for any duplicate numbers.

Pseudo-Random Number Generation
When we get a unique number we are ready to put the new filename together. We have three variables that together make up the filename: $BASE_FN, $DATE_ST, and $RN. The next command puts the pieces together and assigns the filename to the variable to the UNIQUE_FN variable.
UNIQUE_FN=${BASE_FN}.${DATE_ST}.$RN

541

Notice the use of the curly braces ({}) around the first two variables, BASE_FN and DATE_ST. The curly braces are needed because there is a character that is not part of the variable name without a space. The curly braces separate the variable from the character to ensure that we do not get unpredictable output. Because the last variable, $RN, does not have any character next to its name, the curly braces are not needed, but it is not a mistake to add them. The only thing left is to assign the $CURRENT_SECOND value to the LAST_SECOND value and to execute the my_program function, which actually uses the newly created filename. I have commented out the code that would stop the script’s execution. You will need to edit this script and make it suitable for your particular purpose. The mk_unique_filename.ksh shell script is in action in Listing 21.11. yogi@/scripts# ./mk_unique_filename.ksh /tmp/myfilename /tmp/myfilename.120601.131507.03038 /tmp/myfilename.120601.131507.15593 /tmp/myfilename.120601.131507.11760 /tmp/myfilename.120601.131508.08374 /tmp/myfilename.120601.131508.01926 /tmp/myfilename.120601.131508.07238 /tmp/myfilename.120601.131509.07554 /tmp/myfilename.120601.131509.12343 /tmp/myfilename.120601.131510.08496 /tmp/myfilename.120601.131510.18285 /tmp/myfilename.120601.131510.18895 /tmp/myfilename.120601.131511.16618 /tmp/myfilename.120601.131511.30612 /tmp/myfilename.120601.131511.16865 /tmp/myfilename.120601.131512.01134 /tmp/myfilename.120601.131512.19362 /tmp/myfilename.120601.131512.04287 /tmp/myfilename.120601.131513.10616 /tmp/myfilename.120601.131513.08707 /tmp/myfilename.120601.131513.27006 /tmp/myfilename.120601.131514.15899 /tmp/myfilename.120601.131514.18913 /tmp/myfilename.120601.131515.27120 /tmp/myfilename.120601.131515.23639 /tmp/myfilename.120601.131515.13096 /tmp/myfilename.120601.131516.19111 /tmp/myfilename.120601.131516.05964

Listing 21.11

mk_unique_filename.ksh shell script in action. (continues)

Free & Share & Open

542

Chapter 21

/tmp/myfilename.120601.131516.07809 /tmp/myfilename.120601.131524.03831 /tmp/myfilename.120601.131524.21628 /tmp/myfilename.120601.131524.19801 /tmp/myfilename.120601.131518.13556 /tmp/myfilename.120601.131518.24618 /tmp/myfilename.120601.131518.12763 # Listing of newly created files yogi@/tmp# ls -ltr /tmp/myfilename.* -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131507.15593 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131507.03038 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131508.08374 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131508.01926 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131507.11760 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131509.12343 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131509.07554 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131508.07238 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131510.18285 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131510.08496 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131511.30612 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131511.16618 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131510.18895 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131512.19362 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131512.01134 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131511.16865 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131513.10616 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131513.08707

Listing 21.11

mk_unique_filename.ksh shell script in action. (continued)

Pseudo-Random Number Generation

543

-rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131512.04287 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131514.18913 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131514.15899 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131513.27006 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131515.27120 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131515.23639 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131515.13096 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131516.19111 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131516.05964 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131524.21628 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131524.03831 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131516.07809 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131518.24618 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131518.13556 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131524.19801 -rw-r--r-- root system Dec 06 13:15 /tmp/myfilename.120601.131518.12763

Listing 21.11

mk_unique_filename.ksh shell script in action. (continued)

Summary
In this chapter we stepped through some different techniques of creating pseudorandom numbers and then used this knowledge to create unique filenames. Of course these numbers are not suitable for any security-related projects because of the predictability and cyclical nature of computer generated numbers using the RANDOM variable. Play around with these shell scripts and functions and modify them for your

Free & Share & Open

544

Chapter 21 needs. In Chapter 10 we used pseudo-random numbers to create pseudo-random passwords. If you have not already studied Chapter 10, I suggest that you break out of sequence and study this chapter next. In the next chapter we move into a little floating point mathematics and introduce you to the bc utility. Floating point math is not difficult if you use some rather simple techniques. Of course you can make mathematics as difficult as you wish. I hope you gained a lot of knowledge in this chapter and I will see you in the next chapter!

CHAPTER

22
Floating-Point Math and the bc Utility

Have you ever had a need to do some floating-point math in a shell script? If the answer is yes, then you’re in luck. On Unix machines there is a utility called bc that is an interpreter for arbitrary-precision arithmetic language. The bc command is an interactive program that provides arbitrary-precision arithmetic. You can start an interactive bc session by typing bc on the command line. Once in the session you can enter most complex arithmetic expressions as you would in a calculator. The bc utility can handle more than I can cover in this chapter, so we are going to keep the scope limited to simple floating-point math in shell scripts. In this chapter we are going to create shell scripts that add, subtract, multiply, divide, and average a list of numbers. With each of these shell scripts the user has the option of specifying a scale, which is the number of significant digits to the right of the decimal point. If no scale is specified, then an integer value is given in the result. Because the bc utility is an interactive program, we are going to use a here document to supply input to the interactive bc program. We will cover using a here document in detail throughout this chapter.

Syntax
By now you know the routine: We need to know the syntax before we can create a shell script. Depending on what we are doing we need to create a mathematical statement to

545

Free & Share & Open

546

Chapter 22 present to bc for a here document to work. A here document works kind of like a label in other programming languages. The syntax that we are going to use in this chapter will have the following form:
VARIABLE=$(bc \c” read MESSAGE for NODE in $WINLIST do echo “$MESSAGE” | smbclient -M $NODE done

Listing 25.1 Code segment to broadcast a message. (continued)

In the code segment in Listing 25.1 we first define the list file containing the nodes (or users) for which the message is intended. After the node list file is defined we load the file’s contents into the WINLIST variable. We want to give the user the ability to comment out entries in the $WINNODEFILE with a pound sign (#). We also want the user to be able to make comments in the list file after the node/user name. With this increased flexibility we added some filtering in the WINLIST variable assignment. Notice in this assignment that we used grep and awk to do the filtering. First comes the grep command. In this statement we have the entry: grep -v ^#

The -v tells the grep command to list everything except what grep is pattern matching on. The ^# is the notation for begins with a #. The caret (^) is a nice little option that lets us do filtering on lines of data that begin with the specified pattern. To ignore blank lines in a file use the cat $FILE | grep -v ^$ command statement. Also notice the use of the uniq command. This command removes any duplicate line in the file. Any time you need to remove exact duplicate entries you can pipe the output to the uniq command. In the next section we prompt the user for the message to send and read the entire message into the MESSAGE variable. Because we are using a variable for the message the length can not exceed 2048 characters. The smbclient command will truncate the text string to 1600 characters, which should be more than enough for a pop-up message. Now that we have the message and the destination nodes/users, we are ready to loop through each destination in the $WINLIST using the for loop. Usually the smbclient command is an interactive program. The method that we use to supply the message is to echo the $MESSAGE and pipe the output to the smbclient command. The full command statement for sending the message is shown here: echo “$MESSAGE” | smbclient -M $NODE

The -M switch expects a NetBios node name, which is a typical Windows protocol.

Sending Pop-Up Messages from Unix to Windows

623

Adding Groups to the Basic Code
The code segment in Listing 25.1 forms the basis for the entire shell script. We are going to build on the base code to allow us to send messages to specific groups of users by defining the GROUPLIST variable. Each group that is added to the group list is a variable in itself that points to a filename that contains a list of nodes/users, just like the WINNODEFILE variable. By adding this new ability we need a way to tell the shell script that we want the message sent to a particular group. This is where we need to use the getopts command to parse the command line for command switches and switch-arguments. We have used getopts in other chapters in this book so we will get to the details in a moment. There are three steps in defining a group for this shell script. The first step is to add the new group to the GROUPLIST variable assignment statement, which is toward the top of the script. For this example we are adding three groups: UNIX, DBA, and APP-A. The first step looks like the statement shown here.
GROUPLIST=”UNIX DBA APP-A”

The second step is to define a filename for each newly defined group. I like to define a variable to point to the top-level directory, which is /usr/local/bin on my machines. This method makes moving the location of the list files easy with a one-line edit. The code segment is shown here.
GRP_DIR=”/usr/local/bin” UNIX=”${GRP_DIR}/UNIXlist” DBA=”${GRP_DIR}/DBAlist” APP-A=”${GRP_DIR}/APPAlist”

Notice the use of the curly braces (${VAR}) in this code segment. The curly braces are used to separate the variable from the next character if there is no space between the variable name and the next character. The third and final step is to create each of the files and enter the destination nodes in the file with one entry on each line. The code in this shell script allows for you to comment out entries with a pound sign (#) and to add comments after the node/user definition in the file. To use a group the user must specify one or more groups on the command line with the -G switch, followed by one or more groups that are defined in the script. If more than one group is specified, then the group list must be enclosed in double quotes. To send a message to everyone in the Unix and DBA groups use the following command:
# broadcast.ksh -G “UNIX DBA”

Adding the Ability to Specify Destinations Individually
With the code described thus far we are restricted to the users/nodes that are defined in the list files that we created. Now let’s add the ability for a user to specify one or

Free & Share & Open

624

Chapter 25 more message destinations on the command line or by prompting the user for the destination list. These two options require more command-line switches and, in one case, a switch-argument. We are going to add the following command-line switches to this script: -M, -m Prompts the user for the message destination(s) and the message. -H, -h, -N, -n Expects a destination list as a switch-argument. Each switch does the same thing here. The first switch, -M and -m, is the message switch. There is not a switch-argument for this switch, but instead the user is prompted to enter one or more destination nodes/users. The second set of switches each performs the exact same task, and a switch-argument is required, which is a list of destination nodes/users. Some people think of these destination machines as hosts, so I added the -h and -H switches. Other people think of the destination machines as nodes, so I added the -n and -N switches. This way both sets of users can have it their way.

Using getopts to Parse the Command Line
Now we have a bunch of command-line switches, and some of these switches require a switch-argument. This is a job for getopts! As we have studied before, the getopts command is used in a while loop statement. Within the while loop there is a case statement that allows us to take some useful action when a command-line switch is encountered. Whenever a switch is encountered that requires a switch-argument, the argument that is found is assigned to the $OPTARG variable. This $OPTARG is a variable that is build into the getopts command. Let’s look at the getopts command statement and the code segment with the enclosed case statement in Listing 25.2.
# Parse the command-line arguments for any switches. A command# line switch must begin with a hyphen (-). # A colon (:) AFTER a variable (below) means that the switch # must have a switch-argument on the command line while getopts “:mMh:H:n:N:g:G:” ARGUMENT do case $ARGUMENT in echo “\nEnter One or More Nodes to Send This Message:” echo “\nPress ENTER when finished \n\n” echo “Node List ==> \c” read WINLIST ;; h|H|n|N) WINLIST=$OPTARG ;; g|G) GROUP=$OPTARG # $OPTARG is the value of the switch-argument! m|M)

Listing 25.2 Using getopts to parse the command-line switches.

Sending Pop-Up Messages from Unix to Windows

625

# Make sure that the group has been defined for G in $GROUP do echo “$GROUPLIST” | grep -q $G || group_error $G done # All of the groups are valid if you get here! WINLIST= # NULL out the WINLIST variable # Loop through each group in the $GROUP # and build a list of nodes to send the message to. for GRP in $GROUP do # Use “eval” to show what a variable is pointing to! # Make sure that each group has a non-empty list file if [ -s $(eval echo \$”$GRP”) ] then WINLIST=”$WINLIST $(eval cat \$”$GRP” |grep -v ^# \ | awk ‘{print $1}’)” else group_file_error $(eval echo \$”$GRP”) fi done ;; echo “\nERROR: Invalid Augument(s)!” usage exit 1 ;;

\?)

esac done

Listing 25.2 Using getopts to parse the command-line switches. (continued)

Don’t run away yet! The code segment in Listing 25.2 is not too hard to understand when it is explained. In the getopts statement, shown here, we define the valid switches and which switches require a switch-argument and which ones have a meaning without a switch-argument. while getopts “:mMh:H:n:N:g:G:” ARGUMENT

In this getopts statement the switch definitions list, “:mMh:H:n:N:g:G:”, begins with a colon (:). This first colon has a special meaning. If an undefined switch is encountered, which must begin with a hyphen (-), the undefined switch causes a question mark (?) to be assigned to the ARGUMENT variable (you can use any variable name here). This is the mechanism that finds the switch errors entered on the command line.

Free & Share & Open

626

Chapter 25
In the getopts statement the -m and -M switches do not have a switch argument and the -h, -H, -n, -N, -g, and -G switches do require a switch-argument. Whether or not a switch requires an argument is determined by the placement of colons in the definition statement. If a colon (:) appears after the switch in the definition, then that switch requires a switch-argument; if a switch does not have a colon after the switch definition, then the switch does not have a switch-argument. This is really all there is to using the getopts command. Inside the while loop we have an embedded case statement. It is in the case statement that we do something useful with the command-line arguments that are switches. Just remember, getopts does not care what is on the command line unless it has a hyphen (-). This is why we need to test for valid arguments supplied on the command line. In our case statement in Listing 25.2 we take action or make assignments when a valid switch is encountered. When a -M, or -m, switch is found we prompt the user for a list of one or more destination nodes to send the message. When a -h, -H, -n, or -N switch is found, we assign the $OPTARG variable to the WINLIST, which is a list of target users/nodes. When getopts finds -g, or -G, we assign the $OPTARG variable to the GROUP variable. When an undefined switch is found, a question mark (?) is assigned to the ARGUMENT variable. In this situation we give the user an ERROR message, show the usage function, and exit the shell script with a return code of 1, one. Using the eval Function with Variables Let’s go back to the GROUP variable in Listing 25.2 for a minute. Remember that we can have group names assigned to the GROUPLIST variable. Each group assigned to the GROUPLIST variable must have a filename assigned to it that contains a list of destination machines. Now if you think about this you should notice that we have to work with a variable pointing to another variable, which points to a filename. The file contains the list of destination machines. Just how do we point directly to the filename? This is a job for the eval function. The eval function is a Korn shell built-in, and we use it to solve our little dilemma. The eval function works like this in our code. We have the GROUP variable that is one or more groups that the user entered on the command line as a switch-argument to the -G, or -g, switch. Each group that is assigned to the GROUP variable is a pointer to a filename that holds a list of destination machines. To directly access the filename we have to use the eval function. Let’s look at the code segment that uses the eval function in the getopts loop in Listing 25.3. for GRP in $GROUP do # Use “eval” to show the value of what a variable is pointing # to! Make sure that each group has a nonempty list file if [ -s $(eval echo \$”$GRP”) ] then WINLIST=”$WINLIST $(eval cat \$”$GRP” |grep -v ^# \

Listing 25.3 Using eval to evaluate double pointing variables.

Sending Pop-Up Messages from Unix to Windows

627

| awk ‘{print $1}’ | uniq)” else group_file_error $(eval echo \$”$GRP”) fi done

Listing 25.3 Using eval to evaluate double pointing variables. (continued)

We first start a for loop to process each group assigned to the GROUP variable, which is assigned to the GRP variable on each loop iteration. Inside the for loop we first test to see if the group has a group file assigned and if this file size is greater than zero. To do this we use the following command: if [ -s $(eval echo \$”$GRP”) ]

The command substitution, $(eval echo \$”$GRP”), points directly to the file name of the group. We could also use the command substitution, $(eval echo ‘$’$GRP), to directly access the filename. Both statements produce the same result. This eval statement is saying “tell me what this other variable is pointing to, in this statement.” Notice that we use eval two more times in Listing 25.3. We first use eval to assign the destination machine listed in the list file to the WINLIST variable in the command shown here.
WINLIST=”$WINLIST $(eval cat \$”$GRP” | grep -v ^# \ | awk ‘{print $1}’ | uniq)”

In this case we are listing the file with cat and then using grep and awk to filter the output, and uniq to remove any duplicate entries. The next instance of eval is in the error notification. The group_file_error function requires one argument, the group list filename. In this step we are building a list of destination machines if more than one group was given on the command line.

Testing User Input
For any shell script it is extremely important that the information provided by the user is valid. In the broadcast.ksh shell script we have the opportunity to check a lot of user input. Starting at BEGINNING OF MAIN several tests of data need to be made.

Testing and Prompting for WINLIST Data
The first test of user input is a test to ensure that the WINLIST variable is not empty, or NULL. To make this test we use an until loop to prompt the user for a list of destination nodes if the WINLIST is empty. I created a function called check_for_null_winlist

Free & Share & Open

628

Chapter 25 that is used as the loop criteria for prompting the user for a node list input. This function is shown in Listing 25.4. function check_for_null_winlist { if [[ -z “$WINLIST” && “$WINLIST” = “” ]] then return 1 else return 0 fi }

Listing 25.4 Function to check for a Null WINLIST variable.

The only thing that the check_for_null_winlist function in Listing 23.4 does is return a 1, one, as a return code if the $WINLIST variable is empty, or NULL, and return a 0, zero, if the $WINLIST has data assigned. Using this function as the loop criteria in an until loop is easy to do, as shown in the code segment in Listing 25.5.
# Ensure that at least one node is defined to send the message. # If not stay in this loop until one or more nodes are entered # on the command line until check_for_null_winlist do echo “\n\nEnter One or More Nodes to Send This Message:” echo “\n Press ENTER when finished \n\n” echo “Node List ==> \c” read WINLIST done

Listing 25.5 Using an until loop with check_for_null_winlist.

This until loop will continue to execute until the user either enters data or presses CTRL-C.

Testing and Prompting for Message Data
Like the WINLIST data, the MESSAGE variable must have at least one character to send as a message, or we need to prompt the user for the message to send. We use the same type of technique as we did for the WINLIST data. We created the check_for_null _message function to test the MESSAGE variable to ensure that it is not empty, or

Sending Pop-Up Messages from Unix to Windows
NULL. This function returns a 1, one, if the MESSAGE variable is empty and returns a 0, zero, if the MESSAGE variable has data. Check out the function in Listing 25.6. function check_for_null_message { if [[ -z “$MESSAGE” && “$MESSAGE” = “” ]] then return 1 else return 0 fi }

629

Listing 25.6 Function to check for a Null MESSAGE variable.

Using the check_for_null_message function in Listing 25.6 we can execute an until loop until the MESSAGE variable has at least one character. The loop exits when the function returns a 0, zero, for a return code. Look at the until loop in the code segment shown in Listing 25.7.
# Prompt the user for a message to send. Loop until the # user has entered at least one character for the message # to send. until check_for_null_message do echo “\nEnter the message to send:” echo “\nPress ENTER when finished\n\n” echo “Message ==> \c” read MESSAGE done

Listing 25.7 Using an until loop with check_for_null_message.

If the MESSAGE variable already has data assigned, then the until loop will not prompt the user for any input. This is just a test to look for at least one character of data in the $MESSAGE variable.

Sending the Message
At this point we have validated that we have a list of one or more nodes/users to send the message and that the message is at least one character long. As stated before, the $MESSAGE will be truncated at 1600 characters (1600 bytes), which should not be an issue for a pop-up message. If the message is long, then an email is more appropriate.

Free & Share & Open

630

Chapter 25
We have already seen the basics of sending a message with the smbclient command, which is part of the Samba suite of programs. We are going to use the same technique here to send the message. Now we have the list of destination nodes assigned to the WINLIST variable. Let’s look at the code segment to send the message in Listing 25.8. echo “\nSending the Message...\n” # Loop through each host in the $WINLIST and send the pop-up message for NODE in $WINLIST do echo “Sending to ==> $NODE” echo $MESSAGE | $SMBCLIENT -M $NODE # 1>/dev/null if (($? == 0)) then echo “Sent OK ==> $NODE” else echo “FAILED to ==> $NODE Failed” fi done echo “\n”

Listing 25.8 Code segment to send a message to a list of nodes.

We added a few lines of code to the for loop in Listing 25.8. Notice on each loop iteration that the user is informed of the destination for the current loop iteration. When we send the message using the smbclient command we check the return code to see if the message was sent successfully. A 0, zero, return code does not guarantee that the target machine received the message. For example, if the target is a Windows 95 machine and winpopup is not running, then the message is lost and no error message is received back to let you know that the message was not displayed. You will receive a nonzero return code if the machine is not powered up or if the destination machinename cannot be resolved. Also notice the commented-out redirection to /dev/null, after the smbclient command statement. This output redirection to the bit bucket is commented out so that the user can see the result of sending each message. If there is a problem sending a message, then the smbclient event notifications provide better information than a return code for the smbclient command itself. If you want to hide this connection information, uncomment this redirection to the bit bucket.

Putting It All Together
Now that we have covered most of the individual pieces that make up the broadcast.ksh shell script, let’s look at the whole shell script and see how the pieces fit together. The entire shell script is shown in Listing 25.9. Please pay particular attention to the boldface text.

Sending Pop-Up Messages from Unix to Windows

631

#!/bin/ksh # # SCRIPT: broadcast.ksh # AUTHOR: Randy Michael # Systems Administrator # DATE: 1/12/2000 # REV: 1.2.P # PLATFORM: Not platform dependent but requires Samba # # PURPOSE: This script is used to broadcast a pop-up message to # Windows desktops. The Windows machines must be defined in # the $WINNODEFILE file, which is where the master list of # nodes is defined. The $WINNODELIST filename is defined in the # variable definitions section of this shell script. # # You also have the ability of setting up individual GROUPS of # users/nodes by defining the group name to the GROUPLIST variable. # Then define the filename of the group. For example, to define a # Unix and DBA group the following entries need to be made in this # shell script: # # GROUPLIST=”UNIX DBA” # UNIX=”/scripts/UNIXlist” # DBA=”/scripts/DBAlist” # # Assuming that the filenames presented above are acceptable to you. # # There are four options for sending a message: # # 1) Execute this script without any argument prompts for the # message to send and then send the message to all nodes # defined in the $WINNODEFILE. # 2) Specify the “-M” switch if you want to send a message to a # specific node or a list of nodes. The user is prompted for # the message to send. # 3) Specify the -N or -H switches to specify the specific nodes # to receive the message. Add the node list after the -N or # -H switch. # 4) Specify the -G switch, followed by the group name, that the # message is intended be sent. # # EXAMPLES: # To send a message to all nodes defined in the $WINNODEFILE: # # # broadcast.ksh # # To send a message to only the “booboo” and “yogi” machines: # # # broadcast.ksh -H “booboo yogi”

Listing 25.9 broadcast.ksh shell script listing. (continues)

Free & Share & Open

632

Chapter 25

# OR # # broadcast.ksh -N “booboo yogi” # # To send a message to specific machines without specifying # each one on the command line: # # # broadcast.ksh -M # # To send a message to all users in the Unix and DBA # groups only: # # # broadcast.ksh -G “UNIX DBA” # # Each switch is valid in uppercase or lowercase. # # NOTE: This script uses SAMBA!!! SAMBA must be installed # and configured on this system for this shell script # to function! # # EXIT CODES: 0 ==> Normal Execution # 1 ==> Usage Error # 2 ==> Missing Node List File # 3 ==> The “smbclient” program is not in the $PATH # 4 ==> The “smbclient” program is not executable # # REV LIST: # # # set -x # Uncomment to debug this script # set -n # Uncomment to check syntax without any execution # ############################################################ ####### DEFINE BROADCAST GROUPS AND GROUP FILES HERE ####### ############################################################ # Define the file directory for this shell script. GRP_DIR=”/usr/local/bin” # Define all valid groups to send messages GROUPLIST=”UNIX SAP ORACLE DBA APPA APPB” # Define all of the Group files UNIX=”${GRP_DIR}/Unixlist” SAP=”${GRP_DIR}/SAPlist”

Listing 25.9 broadcast.ksh shell script listing.

Sending Pop-Up Messages from Unix to Windows

633

ORACLE=”${GRP_DIR}/ORACLElist” DBA=”${GRP_DIR}/DBAlist” APPA=”${GRP_DIR}/APPAlist” APPB=”${GRP_DIR}/APPBlist” # File that contains the master list of nodes WINNODEFILE=”${GRP_DIR}/WINlist” ############################################################ ################# DEFINE FUNCTIONS HERE #################### ############################################################ function display_listfile_error { # The function is used to inform the users that the # $WINNODEFILE file does not exist. The $WINNODEFILE # filename is defined in the main body of the shell script. echo echo echo echo echo “\n\tERROR: ...MISSING NODE LIST FILE...” “\nCannot find the $WINNODEFILE node list file!” “\nThe $WINNODEFILE file is a file that contains a list of” “nodes to broadcast a message. Create this file with” “one node name per line and save the file.\n\n”

exit 2 } ############################################################ function usage { echo “\nUSAGE: $THISSCRIPT [-M] [-H Host List] [-N Node List] \ [-G Group List]\n\n” echo “EXAMPLES:” echo “\nTo send a message to all nodes defined in the master list” echo “$WINNODEFILE file enter the scriptname without any options:” echo “\n$THISSCRIPT” echo “\nTo send a message to one or more nodes only,” echo “enter the following command:” echo “\n$THISSCRIPT -M” echo “\nTo specify the nodes to send the message to on” echo “the command-line enter the following command:” echo “\n$THISSCRIPT -H \”yogi booboo dino\” “ echo “\nTo send a message to one or more groups use the” echo “following command syntax:” echo “\n$THISSCRIPT -G \”UNIX DBA\” \n\n”

Listing 25.9 broadcast.ksh shell script listing. (continues)

Free & Share & Open

634

Chapter 25

} ############################################################ function check_for_null_message { if [[ -z “$MESSAGE” && “$MESSAGE” = “” ]] then return 1 else return 0 fi } ############################################################ function check_for_null_winlist { if [[ -z “$WINLIST” && “$WINLIST” = “” ]] then return 1 else return 0 fi } ############################################################ function group_error { (($# != 1)) && (echo “ERROR: function group_error expects \ an argument”; exit 1) GRP=$1 echo “\nERROR: Undefined Group - $GRP” usage exit 1 } ############################################################ function group_file_error { (($# != 1)) && (echo “ERROR: function group_file_error expects \ an argument”; exit 1) GPF=$1 echo “\nERROR: Missing group file - $GPF\n” usage

Listing 25.9 broadcast.ksh shell script listing.

Sending Pop-Up Messages from Unix to Windows

635

exit 1 } ############################################################ function check_for_smbclient_command { # Check to ensure that the “smbclient” command is in the $PATH SMBCLIENT=$(which smbclient) # # # # If the $SMBCLIENT variable begins with “which:” or “no” for Solaris and HP-UX then the command is not in the $PATH on this system. A correct result would be something like: “/usr/local/bin/smbclient” or “/usr/bin/smbclient”. ]] || \

if [[ $(echo $SMBCLIENT | awk ‘{print $1}’) = ‘which:’ [[ $(echo $SMBCLIENT | awk ‘{print $1}’) = ‘no’ ]] then echo “\n\nERROR: This script requires Samba to be and configure. Specifically, this script requires that \”sbmclient\” program is in the \$PATH. Please correct and send your message again.\n” echo “\n\t...EXITING...\n” exit elif [ ! then echo echo exit fi } 3 -x $SMBCLIENT ]

installed the this problem

“\nERROR: $SMBCLIENT command is not executable\n” “Please correct this problem and try again\n” 4

############################################################ ################ DEFINE VARIABLES HERE ##################### ############################################################ THISSCRIPT=$(basename $0) # The name of this shell script MESSAGE= WINLIST= # Initialize the MESSAGE variable to NULL # Initialize the list of node to NULL

############################################################ ################ TEST USER INPUT HERE ###################### ############################################################

Listing 25.9 broadcast.ksh shell script listing. (continues)

Free & Share & Open

636

Chapter 25

# Check for the “smbclient” command’s existence check_for_smbclient_command # If no command-line arguments are present then test for # the master $WINNODEFILE, which is defined at the top # of this shell script. if (($# == 0)) # No command-line arguments - Use the master list then [ -s $WINNODEFILE ] || display_listfile_error # Load the file data into the WINLIST variable ignoring # any line in the file that begins with a # sign. WINLIST=$(cat $WINNODEFILE | grep -v ^# \ | awk ‘{print $1}’ | uniq) else # Parse the command-line arguments for any switches. A command # line switch must begin with a hyphen (-). # A colon (:) AFTER a variable (below) means that the switch # must have a switch-argument on the command line while getopts “:mMh:H:n:N:g:G:” ARGUMENT do case $ARGUMENT in echo “\nEnter One or More Nodes to Send This Message:” echo “\nPress ENTER when finished \n\n” echo “Node List ==> \c” read WINLIST ;; h|H|n|N) WINLIST=$OPTARG ;; g|G) GROUP=$OPTARG # $OPTARG is the value of # the switch-argument! # Make sure that the group has been defined for G in $GROUP do echo “$GROUPLIST” | grep -q $G || group_error $G done # All of the groups are valid if you get here! WINLIST= # NULL out the WINLIST variable # Loop through each group in the $GROUP m|M)

Listing 25.9 broadcast.ksh shell script listing.

Sending Pop-Up Messages from Unix to Windows

637

# and build a list of nodes to send the message to. for GRP in $GROUP do # Use “eval” to show what a variable is pointing to! # Make sure that each group has a non-empty list # file if [ -s $(eval echo \$”$GRP”) ] then WINLIST=”$WINLIST $(eval cat \$”$GRP” \ | grep -v ^# | awk ‘{print $1}’ \ | uniq)” else group_file_error $(eval echo \$”$GRP”) fi done ;; echo “\nERROR: Invalid Argument(s)!” usage exit 1 ;;

\?)

esac done ############################################################ ################## BEGINNING OF MAIN ####################### ############################################################ # Ensure that at least one node is defined to send the message. # If not stay in this loop until one or more nodes are entered # on the command line until check_for_null_winlist do echo “\n\nEnter One or More Nodes to Send This Message:” echo “\n Press ENTER when finished \n\n” echo “Node List ==> \c” read WINLIST done ############################################################

fi

# End of “if (($# == 0))” test.

# Prompt the user for a message to send. Loop until the

Listing 25.9 broadcast.ksh shell script listing. (continues)

Free & Share & Open

638

Chapter 25

# user has entered at least one character for the message # to send. until check_for_null_message do echo “\nEnter the message to send:” echo “\nPress ENTER when finished\n\n” echo “Message ==> \c” read MESSAGE done ############################################################ # Inform the user of the host list this message is sent to... echo “\nSending message to the following hosts:\n” echo “\nWIN_HOSTS:\n$WINLIST\n\n” ############################################################ echo “\nSending the Message...\n” # Loop through each host in the $WINLIST and send the pop-up message for NODE in $WINLIST do echo “Sending to ==> $NODE” echo $MESSAGE | $SMBCLIENT -M $NODE # 1>/dev/null if (($? == 0)) then echo “Sent OK ==> $NODE” else echo “FAILED to ==> $NODE Failed” fi done echo “\n” ############################################################ # # This code segment is commented out by default # # Send the message to the Unix machines too using “wall” # and “rwall” if you desire to do so. This code is commented # out by default. # # echo “\nSending Message to the Unix machines...\n” # # echo $MESSAGE | rwall -h $UnixHOSTLIST

Listing 25.9 broadcast.ksh shell script listing.

Sending Pop-Up Messages from Unix to Windows

639

# echo $MESSAGE | wall # echo “\n\nMessage sent...\n\n” # ############################################################ # Remove the message file from the system rm -f $MESSAGE

Listing 25.9 broadcast.ksh shell script listing. (continued)

As you study the script in Listing 25.9 I hope that you can see how the pieces are put together to produce a logical flow. You may have noticed that there is a larger if statement that skips all of the command-line parsing if there are no command-line arguments present. If we do not have anything to parse through, we just use the default master list of machine destinations. I also want to point out a function that is called at the BEGINNING OF MAIN. The check_for_smbclient_command function looks for the smbclient command in the $PATH. Check out this function in Listing 25.10. function check_for_smbclient_command { # Check to ensure that the “smbclient” command is in the $PATH SMBCLIENT=$(which smbclient) # # # # If the $SMBCLIENT variable begins with “which:” or “no” for Solaris and HP-UX then the command is not in the $PATH on this system. A correct result would be something like: “/usr/local/bin/smbclient” or “/usr/bin/smbclient”.

if [[ $(echo $SMBCLIENT | awk ‘{print $1}’) = ‘which:’ ]] || \ [[ $(echo $SMBCLIENT | awk ‘{print $1}’) = ‘no’ ]] then echo “\n\nERROR: This script requires Samba to be installed and configured. Specifically, this script requires that the \”sbmclient\” program is in the \$PATH. Please correct this problem and send your message again.\n” echo “\n\t...EXITING...\n” exit 3 elif [ ! -x $SMBCLIENT ] then

Listing 25.10 check_for_smbclient_command function listing. (continues)

Free & Share & Open

640

Chapter 25

echo “\nERROR: $SMBCLIENT command is not executable\n” echo “Please correct this problem and try again\n” exit 4 fi }

Listing 25.10 check_for_smbclient_command function listing. (continued)

Notice that we use the which command in Listing 25.10 to find the smbclient command in the $PATH. The which command will respond with either the full pathname of the smbclient command or an error message. The two messages look like the following:
# which smbclient /usr/local/samba/bin/smbclient OR # which smbclient which: 0652-141 There is no smbclient in /usr/bin /etc /usr/sbin /usr/ucb /usr/bin/X11 /sbin /usr/local/bin /usr/local/samba/bin /usr/local/bin /usr/dt/bin/ /usr/opt/ifor/ls/os/aix/bin .

If we receive the second message, then the smbclient command cannot be found. Note that this second response begins with which: just before the error code. This is true for AIX and Linux; however, on Solaris and HP-UX the result begins with no as opposed to which:. Using this response we give the user an error message that the smbclient command cannot be found.

Watching the broadcast.ksh Script in Action
You can see the broadcast.ksh shell script in action in Listing 25.11. In this listing we use the -M option to specify that we want to be prompted for both a list of destination machines and a message.

Sending Pop-Up Messages from Unix to Windows

641

[root:yogi]@/scripts# ./broadcast.ksh -M Enter One or More Nodes to Send This Message: Press ENTER when finished Node List ==> booboo Enter the message to send: Press ENTER when finished

Message ==> Please log out at lunch for a system reboot. Sending message to the following hosts:

WIN_HOSTS: booboo

Sending the Message... Sending to ==> booboo added interface ip=10.10.10.1 bcast=10.10.255.255 nmask=255.255.0.0 Connected. Type your message, ending it with a Control-D sent 45 bytes Sent OK ==> booboo

Listing 25.11 broadcast.ksh shell script in action.

My booboo machine is an NT 4 box. The pop-up message that I received is shown in Figure 25.1. The pop-up message in Figure 25.1 is typical for most machines except for Windows 95 and 98. For these two versions of Windows the winpopup program must be running. Most other machines have a similar pop-up message, as shown in Figure 25.1.

Free & Share & Open

642

Chapter 25

Figure 25.1 Pop_Up message sent to a Windows desktop.

Downloading and Installing Samba
You can download the latest version of Samba from the following URL: www .samba.org/samba. From the main page select a download site. Download sites from around the world are available. This page has a link, samba-latest.tar.gx, to the latest version of the source code. If you download the source code you need a C compiler to compile the Samba release. The ./configure file is looking for either gcc or cc when you begin the compilation process. If a suitable C compiler is not found you cannot install the Samba code. For our purposes we can download the available precompiled binary versions of the code. Some of these are back releases, but the smbclient command works just fine. When you download the Samba source code follow these steps to compile the code on your machine. Follow the link to the latest version of Samba. Download the code into a directory on the Unix machine that has plenty of space, at least 500MB. Next, uncompress the release. The code that I downloaded was a tar file that was compressed with gzip, which has a .gz filename extension. Let’s say that you downloaded the Samba code into the /usr/local directory with the filename samba.2.7.latest .tar.gz. You can name it anything you want when you download the file. The following commands in Listing 25.12 are used to uncompress, untar, and install the Samba code.

[root:yogi]@/usr/local > gunzip samba.2.7.latest.tar.gz [root:yogi]@/usr/local > tar -xvf samba.2.7.latest.tar [root:yogi]@/usr/local > cd samba.2.7 [root:yogi]@/usr/local/samba.2.7 > ./configure [root:yogi]@/usr/local/samba.2.7 > make [root:yogi]@/usr/local/samba.2.7 > make install

Listing 25.12 Samba source code installation.

Sending Pop-Up Messages from Unix to Windows
Once the installation is complete you can remove the /usr/local/samba.2.7 directory to regain your disk space. Be aware that your file/directory names and release may differ from the commands shown in Listing 25.12. This source code installation does not create a smb.conf file. In the procedure that is presented in Listing 25.12, the smb.conf file is located in /usr/local/samba/lib/smb.conf. Please refer to the Samba documentation of the release you installed to know where to put this configuration file. For our purposes, and for security, make the file simple! The smbclient command works with a smb.conf file with only a single semicolon, (;). No other entry is required! The semicolon (;) and hash mark (#) are both comment specifications in this file. If you want to use any of the other functionality of Samba you are on your own, and the Samba documentation is your best resource for additional information.

643

Testing the smbclient Program the First Time
Before you start creating the master list file and a bunch of group list files, do a few tests to ensure that you have the correct format, the destination machines are reachable, and the name resolution is resolved for each node. Initially have a list of about five machines. The machines may be referenced in the NetBios world as a machine name or a username. This name resolution varies depending on the Windows network in your environment. My home network does not have NetBios running, so I had to do a little research and I found that there is a file, which coexists with the smb.conf file, that works like a /etc/hosts file. This file is called lmhosts, and you make machine entries into this file just like a regular hosts file, except that the machine-names are entered in uppercase characters. First try the following test. Let’s suppose that I have five users named JohnB, CindySue, Bubba, JonnyLee, and BobbyJoe. For each user in the list we run the following commands. echo echo echo echo echo “Hello “Hello “Hello “Hello “Hello World” World” World” World” World” | | | | | smbclient smbclient smbclient smbclient smbclient -M -M -M -M -M JohnB CindySue Bubba JonnyLee BobbyJoe

Ideally, the response should look something like the following output: added interface Connected. Type sent 13 bytes added interface Connected. Type sent 13 bytes added interface Connected. Type sent 13 bytes added interface ip=10.10.10.1 bcast=10.10.255.255 nmask=255.255.0.0 your message, ending it with a Control-D ip=10.10.10.1 bcast=10.10.255.255 nmask=255.255.0.0 your message, ending it with a Control-D ip=10.10.10.1 bcast=10.10.255.255 nmask=255.255.0.0 your message, ending it with a Control-D ip=10.10.10.1 bcast=10.10.255.255 nmask=255.255.0.0

Free & Share & Open

644

Chapter 25
Connected. Type your message, ending it with a Control-D sent 13 bytes added interface ip=10.10.10.1 bcast=10.10.255.255 nmask=255.255.0.0 Connected. Type your message, ending it with a Control-D sent 13 bytes

If you get responses like the ones shown here, then everything is as we want it to be. If you get output more like the next set of smbclient output, then we have a problem, Houston! added interface ip=10.10.10.1 bcast=10.10.255.255 timeout connecting to 10.10.10.4:139 Error connecting to 10.10.10.4 (Operation already Connection to JohnB failed added interface ip=10.10.10.1 bcast=10.10.255.255 Connection to CindySue failed added interface ip=10.10.10.1 bcast=10.10.255.255 Connection to Bubba failed added interface ip=10.10.10.1 bcast=10.10.255.255 Connection to JonnyLee failed added interface ip=10.10.10.1 bcast=10.10.255.255 Connection to BobbyJoe failed nmask=255.255.0.0 in progress) nmask=255.255.0.0 nmask=255.255.0.0 nmask=255.255.0.0 nmask=255.255.0.0

Notice that the first attempt, to JohnB, timed out on connection. This is good! We know that there is name resolution to this machine but the machine is currently unreachable. I know it is unreachable because I turned the machine off. If a node is not powered up, this is the type of message that we receive. On the other hand, the next four attempts to reach CindySue, Bubba, JonnyLee, and BobbyJoe failed. This is usually an indication that there is no name resolution to get to these machines. When you have this problem, first try to reach the machine by the node name instead of the user name. You can get the name of the machine by leftclicking on the My Computer icon on the Windows desktop. Then select properties. Try the same process of sending the message again, this time using node names. If you still have a problem, consult the Windows Systems Administrators to see if they can help. The other solution is to maintain a lmhosts file, which is a pain to do. The lmhosts file is located in the same directory as the smb.conf file, which is in /usr/local/samba/lib if you downloaded and compiled the distribution from the Samba site. The lmhosts file does not exist by default, so you will have to create the file using the same format as the /etc/hosts file. This problem with this solution is that you have an extra step when you add a node to both the list files for the broadcast.ksh shell script and the lmhosts file.

Other Options to Consider
This is one of those shell scripts that you can do a lot of different things with. Here are a few things that I thought of. Use your imagination, and I’m sure that you can add to this list.

Sending Pop-Up Messages from Unix to Windows

645

Producing Error Notifications
A very good use of this shell script is to set up as many groups as you need to do error notification to users responsible for maintaining particular machines, programs, databases, and applications. When an error is detected in one of the monitoring shell scripts, just send a pop-up message as an immediate notification; the email notification is just gravy on the potatoes. You can make this a powerful tool if you desire.

Add Logging of Unreachable Machines
If you redirect the output of the smbclient command in the shell script to a log file and parse the log file for connection and name resolution errors, you can find out who is not getting some messages, but not all. If a user’s machine is turned off, the message is lost and there is no notification. Even if a message is refused by the host, the return code from the smbclient command is still 0, zero. Keeping a log of the activity and automatically parsing the log after each message is sent can help you find where the rejections occur. Just remember to keep it simple!

Create Two-Way Messaging
I wanted to figure out how to send the message from the Windows machines back to the Unix boxes, but I ran out of time to meet my due date. I am sure that this is not a hard task to solve. This is a good project for you to play around with; I am going to work on this one, too.

Summary
I sure hope that you enjoyed this chapter, and the whole book. The process of writing this book has been a thrill for me. Every time I started a new chapter I had a firm idea of what I wanted to accomplish, but usually along the way I got these little brain storms that help me build on the basic idea that I started with. Some five-page chapters turned into some of the longest chapters in the book. In every case, though, I always tried to hit the scripting techniques from a different angle. Sometimes this resulted in a long script or roundabout way of accomplishing the task. I really did do this on purpose. There is always more than one way to solve a challenge in Unix, and I always aimed to make each chapter different and interesting. I appreciate that you bought this book, and in return I hope I have given you valuable knowledge and insight into solving any problem that comes along. Now you can really say that the solution to any challenge is intuitively obvious! Thank you for reading, and best regards.

Free & Share & Open

APPENDIX

A
What’s on the Web Site
This Appendix shows a list of the shell scripts and functions that are include on the Web site. Each of the shell scripts and functions has a brief description of the purpose.

Shell Scripts
Chapter 2
12_ways_to_parse.ksh:

This script shows the different ways of reading a file line by line. Again there is not just one way to read a file line by line and some are faster than others and some are more intuitive than others. mk_large_file.ksh: This script is used to create a text file that is has a specified number of lines that is specified on the command line.

Chapter 3
No shell scripts to list in Chapter 3.

647

Free & Share & Open

648

Appendix A

Chapter 4 rotate.ksh: This shell script is used as a progress indicator with the appearance of a rotating line. countdown.ksh: This shell script is used as a progress indicator with a countdown to zero.

Chapter 5 fs_mon_AIX.ksh: This shell script is used to monitor an AIX system for full filesystems using the percentage method. fs_mon_AIX_MBFREE.ksh: This shell script is used to monitor an AIX system for full filesystems using the MB free method. fs_mon_AIX_MBFREE_excep.ksh: This shell script is used to monitor an AIX system for full filesystems using the MB free method with exceptions capability. fs_mon_AIX_PC_MBFREE.ksh: This shell script is used to monitor an AIX system for full filesystems using the percentage method with exceptions capability. fs_mon_AIX_excep.ksh: Basic AIX filesystem monitoring using the percent method with exceptions capability. fs_mon_ALL_OS.ksh: This shell script auto detects the UNIX flavor and monitors the filesystems using both percent and MB free techniques with an auto detection to switch between methods. fs_mon_HPUX.ksh: This shell script is used to monitor a HP-UX system for full filesystems using the percentage method.

What’s On the Web Site fs_mon_HPUX_MBFREE.ksh: 649

This shell script is used to monitor an HP-UX system for full filesystems using the MB free method. fs_mon_HPUX_MBFREE_excep.ksh: This shell script is used to monitor an HP-UX system for full filesystems using the percentage method with exceptions capability. fs_mon_HPUX_PC_MBFREE.ksh: This shell script is used to monitor an HP-UX system for full filesystems using the percentage method with exceptions capability. fs_mon_HPUX_excep.ksh: Basic HP-UX filesystem monitoring using the percent method with exceptions capability. fs_mon_LINUX.ksh: This shell script is used to monitor a Linux system for full filesystems using the percentage method. fs_mon_LINUX_MBFREE.ksh: This shell script is used to monitor a Linux system for full filesystems using the MB free method. fs_mon_LINUX_MBFREE_excep.ksh: This shell script is used to monitor a Linux system for full filesystems using the percentage method with exceptions capability. fs_mon_LINUX_PC_MBFREE.ksh: This shell script is used to monitor a Linux system for full filesystems using the percentage method with exceptions capability. fs_mon_LINUX_excep.ksh: Basic Linux filesystem monitoring using the percent method with exceptions capability. fs_mon_SUNOS.ksh: This shell script is used to monitor a SunOS system for full filesystems using the percentage method.

Free & Share & Open

650

Appendix A fs_mon_SUNOS_MBFREE.ksh: This shell script is used to monitor a SunOS system for full filesystems using the MB free method. fs_mon_SUNOS_MBFREE_excep.ksh: This shell script is used to monitor a SunOS system for full filesystems using the percentage method with exceptions capability. fs_mon_SUNOS_PC_MBFREE.ksh: This shell script is used to monitor a SunOS system for full filesystems using the percentage method with exceptions capability. fs_mon_SUNOS_excep.ksh: Basic SunOS filesystem monitoring using the percent method with exceptions capability.

Chapter 6
AIX_paging_mon.ksh:

Shell script to monitor AIX paging space.
HP-UX_swap_mon.ksh:

Shell script to monitor HP-UX swap space.
Linux_swap_mon.ksh

Shell script to monitor Linux swap space.
SUN_swap_mon.ksh:

Shell script to monitor SunOS swap space. all-in-one_swapmon.ksh: Shell script to monitor AIX, HP-UX, Linux, and SunOS swap/paging space.

Chapter 7 uptime_loadmon.ksh: System load monitor using the uptime command.

What’s On the Web Site uptime_fieldtest.ksh: 651

Script to test the location of the latest uptime load information as it changes based on time. sar_loadmon.ksh: Load monitor using the sar command. iostat_loadmon.ksh: Load monitor using the iostat command. vmstat_loadmon.ksh: Load monitor using the vmstat command.

Chapter 8 proc_mon.ksh: Process monitor that informs the user when the process ends. proc_wait.ksh: Process monitor that informs the user when the process starts. proc_watch.ksh: Process monitor that monitors a process as it starts and stops. proc_watch_timed.ksh: Process monitor that monitors a process for a user specified amount of time.

Chapter 9
There are no shell scripts to list in Chapter 9.

Chapter 10 mk_passwd.ksh: This shell script is used to create pseudo-random passwords.

Free & Share & Open

652

Appendix A

Chapter 11 stale_LV_mon.ksh: This shell script is used to monitor AIX stale Logical Volumes. stale_PP_mon.ksh: This shell script is used to monitor AIX stale Physical Partitions. stale_VG_PV_LV_PP_mon.ksh: This shell script is used to monitor AIX stale partitions in Volume Groups, Physical Volumes, Logical Volumes, and Physical Partitions.

Chapter 12 pingnodes.ksh: This shell script is used to ping nodes. The operating system can be AIX, HP-UX, Linux, or SunOS.

Chapter 13
AIXsysconfig.ksh:

This shell script is used to gather information about an AIX system’s configuration.

Chapter 14 chpwd_menu.ksh: This shell script uses sudo to allow support personnel to change passwords. sudo-1.6.3p7.tar.gz: This is a tar ball of the sudo source code.

Chapter 15 hgrep.ksh: This shell script works similar to grep except that it shows the entire file with the pattern match highlighted in reverse video.

What’s On the Web Site

653

Chapter 16 enable_AIX_classic.ksh: Enables all AIX “classic” print queues. print_UP_AIX.ksh: Enables all AIX System V printers and queues. print_UP_HP-UX.ksh: Enables all HP-UX System V printers and queues. print_UP_Linux.ksh: Enables all Linux System V printers and queues. printing_only_UP_Linux.ksh: Enables printing on Linux System V printers. queuing_only_UP_Linux.ksh: Enables queuing on Linux System V printers. print_UP_SUN.ksh: Enables all SunOS System V printers and queues.
PQ_all_in_one.ksh:

Enables all printing and queuing on AIX, HP-UX, Linux, and SunOS by auto detecting the UNIX flavor.

Chapter 17 tst_ftp.ksh: Simple FTP automated file transfer test script. get_remote_dir_listing.ksh: Script to get a remote directory listing using FTP. get_ftp_files.ksh: Shell script to retrieve files from a remote machine using FTP.

Free & Share & Open

654

Appendix A put_ftp_files.ksh: Shell script to upload files to a remote machine using FTP. get_remote_dir_listing_pw_var.ksh: Script to get a directory listing from a remote machine using FTP. The passwords are stored in an environment file somewhere on the system, defined in the script. get_ftp_files_pw_var.ksh: Script to retrieve files from a remote machine using FTP. This script gets its password from an environment file somewhere on the system, defined in the script. put_ftp_files_pw_var.ksh: Script to upload files to a remote machine using FTP. This script gets its password from an environment file somewhere on the system, defined in the script.

Chapter 18 findlarge.ksh: This shell script is used to find “large” files. The file size limit is supplied on the command line and the search begins in the current directory.

Chapter 19 broot: Shell script to capture keystrokes of anyone gaining root access. banybody: Shell script to capture keystrokes of any user defined in the shell script. log_keystrokes.ksh: Shell script to log a user’s keystrokes as they type on the keyboard.

Chapter 20
SSAidentify.ksh:

Shell script to control SSA disk subsystem disk identification lights.

What’s On the Web Site

655

Chapter 21 mk_unique_filename.ksh: This shell script creates unique filenames.

Chapter 22 float_add.ksh: Adds a series of floating point numbers together using the bc utility. float_subtract.ksh: Subtracts floating point numbers using the bc utility. float_multiply.ksh: Multiplies a series of floating point numbers together using the bc utility. float_divide.ksh: Divides two floating point numbers using the bc utility. float_average.ksh: Averages a series of floating point numbers using the bc utility.

Chapter 23 mk_swkey.ksh: Shell script to create a software license key using the hexadecimal representation of the IP address. equate_any_base.ksh: Converts numbers between any base. equate_base_2_to_16.ksh: Converts numbers from base 2 to base 16. equate_base_16_to_2.ksh: Converts numbers from base 16 to base 2.

Free & Share & Open

656

Appendix A equate_base_10_to_16.ksh: Converts numbers from base 10 to base 16. equate_base_16_to_10.ksh: Concerts numbers from base 16 to base 2. equate_base_10_to_2.ksh: Converts numbers from base 10 to base 2. equate_base_2_to_10.ksh: Converts numbers from base 2 to base 10. equate_base_10_to_8.ksh: Converts numbers from base 10 to base 8. equate_base_8_to_10.ksh: Converts numbers from base 8 to base 10.

Chapter 24 operations_menu.ksh: Shell script menu for an Operations staff.

Chaper 25 broadcast.ksh: Shell script to send pop-up messages to Windows desktops. This shell script requires Samba to be installed on the UNIX machine.

Functions
Chapter 2
All of the following 12 functions are different methods to process a file line by line. The two fastest methods are tied for first place and are highlighted in boldface text.

What’s On the Web Site while_read_LINE while_read_LINE_bottom cat_while_LINE_line while_line_LINE while_LINE_line_bottom while_LINE_line_cmdsub2 while_LINE_line_bottom_cmdsub2 while_read_LINE_FD while_LINE_line_FD while_LINE_line_cmdsub2_FD while_line_LINE_FD

657

Chapter 3 send_notification: This function is used to send an email notification to a list of email addresses, specified by the MAILLIST variable defined in the main body of the shell script.

Chapter 4 dots: This function is used as a progress indicator showing a series of dots every 10 seconds, or so. rotate: This function is used as a progress indicator showing the appearance of a rotating line.

Chapter 5
There are no functions to list in Chapter 5.

Chapter 6
AIX_paging_mon:

Function to monitor AIX paging space.
HP_UX_swap_mon:

Function to monitor HP-UX swap space.

Free & Share & Open

658

Appendix A
Linux_swap_mon:

Function to monitor Linux swap space.
SUN_swap_mon:

Function to monitor SunOS swap space.

Chapter 7
There are no functions to list in Chapter 7.

Chapter 8
There are no functions to list in Chapter 8.

Chapter 9 check_HTTP_server: This function is used to check an application Web server and application URL pages.

Chapter 10 in_range_random_number: This function creates pseudo-random numbers within one and a “max value”. load_default_keyboard: This function is used to load a USA 102-key board layout into a keyboard file. check_for_and_create_keyboard_file: If the $KEYBOARD_FILE does not exist then ask the user to load the “standard” keyboard layout, which is done with the load_default_keyboard function. build_manager_password_report: Build a file to print for the secure envelope.

Chapter 11
There are no functions to list in Chapter 11.

What’s On the Web Site

659

Chapter 12 ping_host: This function executes the correct ping command based on UNIX, the UNIX flavor, AIX, HP-UX, Linux, or SunOS. ping_nodes: This function is used to ping a list of nodes stored in a file. This function requires the ping_host function.

Chapter 13
All of these functions are used in gathering system information from an AIX system. Refer to Chapter 13 for more details in the AIXsysconfig.ksh shell script. get_host get_OS get_OS_level get_ML_for_AIX get_TZ get_real_mem get_arch get_devices get_long_devdir_listing get_tape_drives get_cdrom get_adapters get_routes get_netstats get_fs_stats get_VGs get_varied_on_VGs get_LV_info get_paging_space get_disk_info get_VG_disk_info get_HACMP_info get_printer_info get_process_info get_sna_info get_udp_x25_procs get_sys_cfg get_long_sys_config get_installed_filesets check_for_broken_filesets last_logins

Free & Share & Open

660

Appendix A

Chapter 14
There are no functions to list in Chapter 14.

Chapter 15
There are no functions to list in Chapter 15.

Chapter 16
AIX_classic_printing:

Enables AIX print queues using the AIX “classic” printer subsystem.
AIX_SYSV_printing:

Enables AIX printers and queues using System V printing subsystem.
HP_UX_printing:

Enables HP-UX printers and print queues using System V printing.
Linux_printing:

Enables Linux printers and print queues using System V printing.
Solaris_printing:

Enables SunOS printers and print queues using System V printing.

Chapter 17 pre_event: Function to allow for pre events before processing. post_event: Function to allow for post events after processing.

Chapter 18
There are no functions to list in Chapter 18.

What’s On the Web Site

661

Chapter 19
There are no functions to list in Chapter 19.

Chapter 20 man_page: Function to create man page type information about the proper usage of the SSAidentify.ksh shell script. twirl: Progress indicator that looks like a “twirling”, or rotating line. all_defined_pdisks: Function that lights all disk identification lights for all defined pdisks. all_varied_on_pdisks: Function that lights all disk identification lights that are in varied-on Volume Groups. list_of_disks: Function that acts on each pdisk by turning on/off the SSA disk identification lights.

Chapter 21 get_random_number: This function produces a pseudo-random between 1 and 32,767. in_range_random_number: Create a pseudo-random number less than or equal to the $UPPER_LIMIT value, which is user defined.

Chapter 22
There are no functions to list in Chapter 22.

Free & Share & Open

662

Appendix A

Chapter 23
There are no functions to list in Chapter 23.

Chapter 24
There are no functions to list in Chapter 24.

Chapter 25 check_for_null_message: Checks to see if a variable is empty. check_for_null_winlist: Checks to see if a variable is empty. check_for_smbclient_command: Checks for the existence of the smbclient command and ensures the file is executable.

Index
SYMBOLS
* (asterisk), 28 \ (backslash), 2, 103, 265 ` (back tic mark), 16 ^ (caret), 105, 207, 622 $? (check return code), 25–26 : (colon) checking NFS for, 113–118 getopts command and, 229, 562, 604 {} (curly braces), 104, 541 . (decimal point), 554 $ (dollar sign) numeric test comparison and, 120 variable name and, 13, 151 && (double ampersands), 407 $$ (double dollar signs), 524 == (double equal signs), 217 || (double pipes), 407 “ (double quotes) multiword string patterns and, 392 uses of, 16 variable and, 115 / (forward slash), 471 ‘ (forward tic mark), 2, 16 # (hash mark), 5 - (hyphen), 562 - (minus) sign, 554 : (no-op), 58 % (percent) character, 103, 526 |& (pipe ampersand), 232 + (plus) sign, 554 # (pound) operator, 527 ? (question mark), 28, 562

A accessing value of $# positional parameter, 45 variable data, 13 adding list of numbers, 547–551, 555 AIX classic printer subsystem, 38, 404–408 df -k command output, 130, 131 iostat command output, 186 Logical Volume Manager (LVM), 298 lsps command, 146–147 paging monitor, 149–155 ping command syntax, 320 pwdadm command, 385–389 sar command output, 188–189 system monitoring, 98–103 system snapshot commands, 338–340 system snapshot listing, 341–351 system snapshot report output, 353–366 System V output, 426 System V printing, 408–414 uptime command output, 180–181 vmstat command output, 191 See also stale disk partition, monitoring for

663

Free & Share & Open

664

Index aliases for /etc/sudoers file, 384, 488 all_defined_pdisks function, 501–503 all_varied_on_pdisks function, 503–505 API (application program interface), 261 application monitoring APIs and SNMP traps, 261 HTTP server, checking, 259–260 local processes, 252–254 Open Secure Shell and, 254–256 Oracle databases, checking for, 256–259 overview of, 251, 260 application program interface (API), 261 arguments, command switch and, 229. See also command-line arguments arithmetic operators modulo, 267, 526 overview of, 17 array creating, 418–419 loading, 46–47, 264, 265–266, 280–281 one-dimensional, 265 uses of, 425 working with, 419–420 array pointer, 264 ASCII text, 2 assigning variable, 13 asterisk (*), 28 at command, 28, 96 auditing root access, 476, 483–486 auto-detect techniques, 118 automated event notification basics of, 79–81 file system monitoring and, 143–144 techniques for, 79 automated FTP file transfer, 39 automated host pinging, 37 averaging series of numbers, 579–582 awk statement, 100, 307 bc utility float_add script, 546–552 float_average script, 579–582 float_divide script, 573–579 float_multiply script, 565–570 float_subtract script, 556–561 functions, creating, 582–583 here document, using, 555–556, 564–565 Linux swap space monitor and, 160–161 math statement, building, 554–555, 563–564 overview of, 545 parsing command-line arguments with getopts, 561–563 parsing command line for valid numbers, 570–572 scale, removing from scripts, 582 scale, setting, 161, 165, 556 Solaris swap space monitor and, 165 syntax, 545–546 testing for integers and floating-point numbers, 552–554 bdf command, 132–133 bin directory, 103 blank line, removing from file, 44 boot logical volume, 300 bounce account, 480 break command, 9 broadcasting message to all users, 621–622 error notifications, 645 groups, adding, 623 to individual destinations, 623–627 log file and, 645 overview of, 43, 619 script for, 631–639 sending message, 629–630 testing user input, 627–629 build_manager_password_report function, 271–274 built-in tests, 26

B background, co-process and, 245–246 backslash (\), 2, 103, 265 back tic mark (`), 16 basename command, 94–95 basename $0 command, 530 bc command and floating-point math, 40–41, 545

C calling function, 2 capturing large list of files, 39 user keystrokes, 40, 475–476, 480–483

Index caret (^), 105, 207, 622 case sensitivity, 1 case statement, 8, 437–438 catching delayed command output, 32–33 cat command, 57–58 cc, 371 CD-ROM files, script stub and, 5–6 monitoring and, 98–99 check_exceptions function, 114 check_for_and_create_keyboard_file function, 270–271 check_for_null_message function, 629 check_for_null_winlist function, 628 check_for_smbclient_command function, 639–640 check_HTTP_server function, 259–260 check return code ($?), 25–26 chmod command, 18–20 cleanup function, 500–501 colon (:) checking NFS for, 113–118 getopts command and, 229, 562, 604 columns heading, removing in command output, 45–46 command-line arguments overview of, 13–14 parsing with getopts, 29–30, 244–245, 561–563, 602–604, 624–626 special parameters and, 15–16 testing and parsing, 275–279 commands at, 28, 96 AIX classic print control, 404–408 AIX System V print control, 408–414 basename, 94–95 basename $0, 530 bc, 40–41, 545 break, 9 cat, 57–58 catching delayed output, 32–33 chmod, 18–20 compress, 493 configure, 371–375 crontab, 27 cut, 150 date, 467, 472, 540 df -k, 98, 130–132 disk subsystem, 298–299 echo, 35, 88, 89, 498–499, 524–525 echo $#, 275 enq -A, 404–405 env, 456 executing in sub-shell, 417 exit, 9, 483 find, 39, 465, 466–473 free (Linux), 148 ftp, 442, 443, 463 gunzip, 371 gzip, 493 hostname, 467, 478 HP-UX print control, 414–417 iostat, 179, 186–188, 203–208 kill, 500–501 last, 22–23 line, 54, 58–62 Linux print control, 417–422 list of, 10–12 lpc (AIX), 408–412 lpc (Linux), 417–418 lpc (Solaris), 425–429 lpstat (AIX), 412–414 lpstat (HP-UX), 414–417 lpstat (Solaris), 429–431 lsdev, 501–503 lslv, 299 lsps (AIX), 146–147 lspv, 299, 304, 504 lsvg, 298, 299, 308 lsvg -o, 504 mail notification, 34–35, 80–83 make, 371, 375–377 make install, 377 manual page, printing, 465 more, 392, 393, 399, 471 number base conversion, 41–42 pg or page, 392, 393, 399 ping, 37–38, 251, 319 printf, 41–42, 586–587 ps, 23 ps auxw, 179, 213–214 ps -ef, 216, 252, 257 pwd, 466, 467, 471 pwdadm, 385–389 read, 53–54, 60 real-time user communication, 24

665

Free & Share & Open

666

Index commands (continued) remote, running, 255–256 removing columns heading in output, 45–46 return, 9 rsh, 20–21 running on remote host, 20–21 rwall, 24 sar, 179, 188–191, 197–203, 214 select, 42–43 sendmail, 34–35, 83–84, 330 set -A, 46–47, 265, 418 shift, 14–15, 571, 607 smbclient, 619–621, 630 Solaris print control, 425–431 sqlplus, 257–258 ssaidentify, 496–497, 503 ssaxlate, 315–316, 496, 504 ssh, 254 sudo program and, 369–370 su (switch user), 478, 492 swapinfo (HP-UX), 147–148 swap (Solaris), 148–149 symbol commands, 13 system snapshot for AIX, 338–340 tail, 45–46, 405 talk, 24 tee, 223 tee -a, 352, 422 time, 56, 67 touch, 446 tput, 38, 389, 400–401 tput rmso, 392 tput smso, 154, 392 tr, 24–25 tty, 223 typeset, 24–25, 41, 529, 585–586 uname, 128, 176 uniq, 43–44, 622 uptime, 179, 180–186, 194–197 user information, 22–23 vmstat, 179, 191–193, 208–212 w, 22 wall, 24 which, 640 who, 22 write, 24 See also getopts command; script command command substitution back tics and, 16 description of, 60 experiment using, 393 options for, 323 timing methods of, 77–78 comments, 4–6 communicating with users, 23–24 compiling sudo, 371–372 compress command, 493 compressing file, 493 configure command, 371–375 configuring sudo, 378–384 continue, 9 control structures, 6–8 converting numbers between bases base 2 (binary) to base 16 (hexadecimal), 587–590 base 8 to base 16, 586 base 10 (decimal) to base 16 (hexadecimal), 590–593 base 10 to base 16, 586 base 10 to hexadecimal, 587 base 10 to octal, 587 beginning of main, 606–608 overview of, 41–42, 585 parsing command-line argument with getopts, 602–604 sanity checks, 604–606 software key, creating, 594–597, 608 translation between any base, 597–608 typeset command syntax, 585–586 co-process with background function, making, 30–32 process monitoring and, 245–246 setting up, 230–231 countdown indicator, 91–96 COUNT variable, 570 CPU load monitoring. See system load monitoring cron table automated hosts pinging and, 335 file system monitoring and, 143–144 overview of, 27–28 pinging and, 321 printing, queuing, and, 409, 418 silent running and, 29 curly braces ({}), 104, 541

Index current directory, searching and, 472 cut command, 150 echo $# command, 275 egrep statement file system monitoring and, 144 grep compared to, 99, 410 -v argument, 99–100, 133–134 email as repository for log files, 479–480, 486, 493 enclosures, 16 encryption key, 254 enq -A command, 404–405 enterprise management tool, 85, 86, 261 env command, 456 EOF character string, 555, 564 error log, 520 error notification, 645 escaping special character, 2, 265 /etc/motd file, 23 /etc/sudoers file samples, 378–384, 486–488 troubleshooting, 494 eval function, 626–627 event notification basics of, 79–81 file system monitoring and, 143 monitoring for stale disk partition, 316 swap space monitoring and, 177 techniques for, 79 events, pre, startup, and/or post, running, 228–229, 249 exceptions capability exceptions file, 103–110 MB of free space with exceptions method, 113–118 print queue and, 439 executing command in sub-shell, 417 shell script recursively, 485–486 exit command, 9, 483 exit criteria, 616 exit signals, 21 exporting password variable, 456

667

D date command, 467, 472, 540 debug mode, automated FTP and, 463 decimal point (.), 554 declaring function, 3 shell, 3–4 default shell, 3 defining function, 120, 340–341, 351 trigger value, 118 /dev/random, 524 df -k command AIX output, 130, 131 Linux output, 131 overview of, 98 SUN/Solaris output, 132 dial-out modem software, 84–85 directory, adding to path, 466 directory listing, saving remote, 444–446 dividing numbers, 573–579 dollar sign ($) numeric test comparison and, 120 variable name and, 13, 151 dotting filename, 456–457 double ampersands (&&), 407 double bracket test for character data, 73 double dollar signs ($$), 524 double equal signs (==), 217 double parentheses mathematical test, 73, 151 double pipes (||), 407 double quotes (“) multiword string patterns and, 392 uses of, 16 variable and, 115 downloading Samba, 642–643 sudo program, 370–371

F E echo command cursor control commands for, 498–499 RANDOM environment variable and, 524–525 series of dots and, 35, 88, 89 file descriptors overview of, 54 parsing file with, 63–66 timing data and, 67, 73 filename creating unique, 535–543 dotting, 456–457

Free & Share & Open

668

Index filename (continued) log files, 478 See also pseudo-random number, creating files aliases for sendmail, 83 capturing large list of, 39 CD-ROM, 5–6 compressing, 493 .forward, 82–83, 480 gzip, 371 highlighting text in, 38 large, finding, 465 permissions and chmod command, 18–20 processing line by line, 33 .profile, ownership of, 477 searching for newly created, 473 See also filename; find command file system monitoring automated execution, 143–144 command syntax, 98–103 egrep statement, modifying, 144 event notification, 143 exceptions capability, adding, 103–110 full, defining, 100–101 MB of free space method, 110–113 MB of free space with exceptions method, 113–118 percentage used-MB free combination, 118–128 techniques for, 97 Unix flavors and, 128–130 See also operating system (OS) File Transfer Protocol. See FTP find command large file script, creating, 466–472 options for searching, 472–473 overview of, 39, 465 syntax, 466 flexibility in scripting, 249 floating-point math float_add script, 546–552 float_average script, 579–582 float_divide script, 573–579 float_multiply script, 565–570 float_subtract script, 556–561 functions, creating, 582–583 here document, using, 555–556, 564–565 Linux swap space monitor and, 160–161 math statement, building, 554–555, 563–564 overview of, 545 parsing command-line arguments with getopts, 561–563 parsing command line for valid numbers, 570–572 scale, removing from scripts, 582 scale, setting, 161, 165, 556 Solaris swap space monitor and, 165 syntax, 545–546 testing for integers and floating-point numbers, 552–554 floating printer, 439 for ... in statement, 7 .forward file, 82–83, 480 forward slash (/), 471 forward tic mark (‘), 2, 16 free command (Linux), 148 FTP (File Transfer Protocol) automation of, 39, 441, 444 controlling execution with command-line switches, 463 debug mode, adding, 463 getting files from remote system, 446–450 here document and, 442–443 log file, adding, 463 modifying script to use password variables, 456–463 pre and post events, 449 replacing hard-coded passwords with variables, 452–456 saving remote directory listing, 444–446 syntax for, 441–444 typical file download, 442 uploading files to remote system, 450–452 ftp command, 442, 443, 463 full pathname, 471 functions all_defined_pdisks, 501–503 all_varied_on_pdisks, 503–505 build_manager_password_report, 271–274 calling, 119 check_exceptions, 114 check_for_and_create_keyboard_file, 270–271 check_for_null_message, 629 check_for_null_winlist, 628 check_for_smbclient_command, 639–640 check_HTTP_server, 259–260

Index cleanup, 500–501 converting shell script into, 175–176 declaring, 3 defining, 120, 340–341, 351 eval, 626–627 form of, 3 get_max, 212–213 getopts, 278 get_random_number, 525 in_range_fixed_length_random_number, 527–528 in_range_random_number, 267–268 as interpreted, 2 list_of_disks, 506–507 load_default_keyboard, 268–270 man_page, 499 mathematical, built-in, 18 overview of, 2–3 positional parameters and, 14 send_notification, 83–84 show_all_instances_status, 257 show_oratab_instances, 256 simple_SQL_query, 258 trap_exit, 275 twirl, 499–500 usage, 274–275, 497–498 on Web site for book, 656–662 uptime field test solution and, 184–186 See also hgrep (highlighted grep) group, broadcasting message to, 623 gunzip command, 371 gzip command, 493 gzip file, 371

669

H hash mark (#), 5 hdisk# cross-referencing to pdisk#, 520 overview of, 495 translating to pdisk#, 496 here document bc utility and, 545–546, 555–556, 564–565 FTP process and, 442–443 swap space monitor and, 161, 165–166 syntax for, 9–10 hgrep (highlighted grep) building shell script, 393–394 listing, 394–399 overview of, 391 reverse video control, 392–393 highlighting text in file, 38 $HOME/.profile, 617 hostname command, 467, 478 HP-UX bdf command output, 132–133 iostat command output, 186–187 ping command syntax, 320 print control commands, 414–417 sar command output, 189 swapinfo command, 147–148 swap space monitor, 155–160 uptime command output, 181–182 vmstat command output, 191 HTTP server, checking, 259–260 hyphen (-), 562

G gcc, 371 get_max function, 212–213 getopts command automating FTP and, 463 limitations of, 246 parsing command-line arguments with, 29–30, 244–245, 278–279, 561–563, 602–604, 624–626 process monitoring and, 218, 228, 229–230 getopts function, 212–213 get_random_number function, 525 global variable, 582 goal of script, 2 grep statement egrep compared to, 99 exceptions capability and, 104–105 process monitoring and, 214 ps -ef command and, 216 rows and, 307

I identifying hardware components, Unix flavor and, 495 SSA disk, 496–497 if statement, tests used in, 330 if ... then ... elif ... (else) statement, 7 if ... then ... else statement, 6 if ... then ... fi statement, 417 if ... then statement, 6

Free & Share & Open

670

Index input redirection, 58 in_range_fixed_length_random_number function, 527–528 in_range_random_number function, 267–268 integer, testing for, 552–554 iostat command, 179, 186–188, 203–208 IP address, creating software key based on, 594–597, 608 automated FTP, 444 averaging list of numbers, 581 base 2 (binary) to base 16 (hexadecimal) conversion, 588–589 base 10 (decimal) to base 16 (hexadecimal) conversion, 590–592 broadcasting message, 621–622 build_manager_password_report function, 272 case statement for iostat fields of data, 188 case statement for sar fields of data, 190 case statement for vmstat fields of data, 192–193 cat $FILENAME|while line LINE method, 60 cat $FILENAME|while LINE=$(line) method, 62 cat $FILENAME|while LINE=`line` method, 59 cat $FILENAME|while read LINE method, 57 check_exceptions function, 114 check_for_and_create_keyboard_file function, 270–271 check_for_null_message function, 629 check_for_null_winlist function, 628 check_for_smbclient_command function, 639–640 check_HTTP_server function, 259–260 cleanup function, 501 configure command output, 371–375 controlling case statement to pick OS, 437–438 co-process, 231 countdown indicator, 92–94 cursor control using echo command, 499 dividend and divisor, extracting, 573–574 equate_any_base, 598–601 /etc/sudoers file samples, 378–381, 381–384, 486–488 exceptions file, 109 exceptions file that worked best with testers, 127–128 filename, creating unique, 536–539, 541–543 file system monitoring for AIX, 101–102 file system monitoring for AIX with exceptions capability, 106–109

J job control, 28–32 junk variable, 420

K kill command, 500–501

L large file, searching for, 466–473 last command, 22–23 $LENGTH, testing for integer value, 277–278 line, rotating, creating, 35–36, 89–91, 95–96, 499–500 line command, 54, 58–62 Linux controlling queuing and printing individually, 422–425 df -k command output, 131 free command, 148 iostat command output, 187 ping command syntax, 320 print control commands, 417–422 sar command output, 189 swap space monitor, 160–164 System V output, 409, 426 uptime command output, 182 vmstat command output, 192 linx command-line browser, 259–260 listings AIX lsps -s data gathering, 150 AIX paging monitor, 151–153, 154 AIX system snapshot commands, 338–340 AIX System V printing, 411–412 all_defined_pdisks function, 503 all-in-one paging and swap space monitor, 169–175

Index finding large file, 467–470, 471–472 fixed-length random number output, 528 float_add script, 547–551, 555 float_average script in action, 581 float_divide script, 574–579 float_multiply script, 565–570, 572 float_subtract script, 556–561, 565 for loop enabling classic AIX print queues, 406-407 FTP file download, 442 full filesystem on yogi script, 110 full filesystem script, 103 getopts command line parsing, 279 getopts command usage, 229–230 getopts function, 278 get_random_number function, 525 get remote directory listing, 445 get remote directory listing, hard-coded passwords removed, 457–458 getting files from remote system, 446–448, 449–450 getting files from remote system, hardcoded passwords removed, 458–460 grep mistake, 104–105 here document for FTP, 442–443 hgrep, 394–399 HP-UX print control, 416 HP-UX swapinfo -tm command output, 155 HP-UX swap space monitor, 157–158 HP-UX swap space report, 159–160 in_range_fixed_length_random_number function, 527–528 in_range_random_number function, 268, 526 iostat load monitoring, 203–205, 207–208 $LENGTH, testing for integer value, 277–278 Linux, controlling printing individually, 422–423 Linux, controlling queuing individually, 424 Linux print control, 420–421 Linux swap space monitor, 162–163, 164 list_of_disks function, 506–507 load_default_keyboard function, 268–269 loading KEYS array, 280–281 logging keystrokes, 480–482 logging root access, 483–485 logic code for large and small filesystem freespace script, 119 looping in background, 88 loop list, building, 281–282 lpstat command output, 413, 429–430 lpstat command using -a and -p, 414, 430 lpstat or enq -A command output, 404 lpstat -W or enq -AW command output, 405 lsdev listing of pdisks, 502 lsvg -l appvg2 rootvg command output, 300 LV, loop to show number of stale PPs from each, 302 LV statistics, 301 mail code segment, 81 mail service, testing, 82 make command output, 375–377 make install command output, 377 MB of free space method, 111–113 MB of free space with exceptions method, 115–118 monitor all OS, 134–141 monitoring administration users, 489–492 monitoring application process, 253 my_sql_query.sql, 257 operating system test, 129 operations menu, 612–616 paging and swap space report, 146 parsing command line, 606–607 parsing command-line switches with getopts, 624–625 parsing numbers from command line, 572 password file with variable exported, 453–454 password file with variable not exported, 454 password report, 273–274 password report, printing, 283–284 password, testing visibility of, 455 percentage free-MB free combination, 121–127 pinging, automated hosts, with notification, 324–328, 331 pop-up messages, sending to Windows, 631–639, 641

671

Free & Share & Open

672

Index listings (continued) printing and queuing all-in-one, 431–436 process monitoring, 218–222, 223 process monitoring and logging, 224–227, 228 process monitoring timed execution, 232–244 process monitoring timed execution in action, 248 process startup loop, 216–217 process wait, 218 progress indicator background function, 89 pseudo-random number, creating, 531–535 pseudo-random password, 284–294 pseudo-random password, building new, 282 PV statistics, 305 reverse video menu options, 610–611 reverse video message bar, 611 rotate function, 90, 91 running remote command, 255–256 running total of numbers, 580–581 Samba source code installation, 642 sar load monitoring, 198–200, 202 script for timing of line by line processing, 55–56 script session, command-line, 476–477 secure shell login, 254–255 sending message to list of nodes, 630 send_notification function, 83–84, 331 shell script starter file, 5–6 shift command, 14–15 show_all_instances_status function, 257 show_oratab_instances function, 256 simple_SQL_query function, 258 software key, creating, 594–595 Solaris print commands, 427–428 Solaris swap space monitor, 166–168, 169 sorted timing data by method, 75–76 SQL+ Oracle query, 258–259 SSA identify, 507–519 stale LV monitoring, 303–304 stale PP monitoring, 305–306 sudo, using first time, 385 sudo, using in shell script, 386–388 sudo log file, 389–390 system snapshot for AIX, 341–351 system snapshot for AIX report output, 353–366 testing command input, 72 testing command-line arguments, 276–277 testing for integers and floating-point numbers, 552–553 timing command substitution methods, 77 timing data for each method, 73–75 timing script, 67–72 twirl function, 500 typeset command in random number function, 529 typeset command to fix length of variable, 529 uploading files to remote system, 450–452 uploading files to remote system, hardcoded passwords removed, 460–462 uptime field test solution, 184–185 uptime load monitoring, 194–196, 197 usage function, 274, 497–498 /usr/local/bin/exceptions file, 142 verifying number base variables, 604–605 VG, LV, and PV monitoring with resync, 308–313 vmstat load monitoring, 208–212 while_line_LINE_Bottom method, 59, 76 while LINE=`line` from Bottom, 61 while LINE=$(line) from Bottom method, 62 while line LINE with file descriptors method, 66 while LINE=$(line) with file descriptors method, 65 while LINE=’line’ with file descriptors method, 64–65 while read $FILENAME from Bottom, 58 while read LINE method, 64 list_of_disks function, 506–507 load_default_keyboard function, 268–270 loading array, 46–47, 264, 265–266, 280–281 default keyboard layout, 268 log file automated FTP and, 463 automated hosts pinging, adding to, 333–334

Index emailing, 479–480, 486, 493 filename, 478 monitoring for stale disk partition, 316 pop-up message and, 645 print queue and, 439 sudo program, 389–390 swap space monitoring and, 177 user activity and, 478–479 logging process starts and stops, 223–228 logical AND, 17, 407 logical OR, 17, 407 Logical Volume Manager (LVM, AIX), 298 loop counter, 410 looping techniques for parsing file line by line cat $FILENAME|while line LINE method, 60–61 cat $FILENAME|while LINE=$(line) method, 61–62 cat $FILENAME|while LINE=`line` method, 59–60 cat $FILENAME|while read LINE method, 57–58, 77 command syntax, 53–54 file for testing timing of, 54–56 techniques for, 33 timing command substitution methods, 77–78 timing data for each method, 73–77 timing methods, 66–67 timing script, 67–72 while_line_LINE_Bottom method, 58–59, 76 while LINE=`line` from Bottom, 61 while LINE=$(line) from Bottom method, 62 while line LINE with file descriptors method, 66 while LINE=$(line) with file descriptors method, 65–66 while LINE=’line’ with file descriptors method, 64–65 while read $FILENAME from Bottom method, 58 while read LINE with file descriptors method, 63–64, 76–77 lpc command AIX, 408–412 Linux, 417–418 Solaris, 425–429 lpstat command AIX, 404–405, 412–414 HP-UX, 414–417 Solaris, 429–431 lsdev command, 501–503 lslv command, 299 lsps command (AIX), 146–147 lspv command, 299, 304, 504 lsvg command, 298, 299, 308 lsvg -o command, 504

673

M mail command automated event notification with, 80–81 syntax, 34 -v switch, 82–83 mail notification techniques, 34–35 mailx command automated event notification with, 80 syntax, 34 -v switch, 82–83 maintenance window communicating with users and, 23 printing and, 439 make command, 371, 375–377 Makefile, configuring, 371–375 make install command, 377 man_page function, 499 manual page, printing, 465 math bc command for floating-point, 40–41, 545 functions, built-in, 18 modulo operator, 267, 526 operators, 17 See also bc utility MB (megabytes), size of, 473 measurement type, 111 memory, paging and swap space and, 145–146 memory leak, 153 menu, creating for Operations staff, 609–618 select command and, 42–43 message, broadcasting. See pop-up messages, sending to Windows message bar, creating, 611–612 mget subcommand, 449 MIB (Management Information Base), 86 Miller, Todd, 370, 378

Free & Share & Open

674

Index minus (-) sign, 554 model dialing software, 84–85 modulo arithmetic operator, 267, 526 monitoring. See application monitoring; file system monitoring; process monitoring; script command; system load monitoring more command, 392, 393, 399, 471 multiplying list of numbers, 565–570, 572 Operations staff, menu for overview of, 609 reverse video syntax, 610–618 operators math, 17 modulo arithmetic, 267, 526 numeric, 120 pound (#), 527 Oracle database, checking for, 256–259 OS. See operating system outbound mail, problems with, 82–84 output control, 28–32

N named pipe, creating, 493 nlist subcommand, 444–446 no-op (:), 58 notification of event. See event notification null value check, 114 null variable, testing for, 44, 115 number. See bc utility; pseudo-random number, creating number base conversion base 2 (binary) to base 16 (hexadecimal), 587–590 base 8 to base 16, 586 base 10 (decimal) to base 16 (hexadecimal), 590–593 base 10 to base 16, 586 base 10 to hexadecimal, 587 base 10 to octal, 587 beginning of main, 606–608 overview of, 41–42, 585 parsing command-line argument with getopts, 602–604 sanity checks, 604–606 software key, creating, 594–597, 608 translation between any base, 597–608 typeset command syntax, 585–586 numeric test comparison, 120

P padding number with leading zeros, 527–530 page command, 392, 393, 399 pager notification, 143 paging space. See swap space parameters positional, 13–14, 45, 601–602 special, 15–16 parsing command-line arguments, 29–30, 275–279 command-line arguments with getopts, 244–245, 561–563, 602–604, 624–626 command line for valid numbers, 570–572 file with file descriptors, 63–66 file with line command, 54, 58–62 See also processing file line by line passwords hard-coded, 446 hard-coded, replacing, 452–456 page of, printing, 264, 271–272, 283–284, 294 password environment file, creating, 456–457 pwdadm command, 385–389 randomness and, 263 root, auditing, 476, 483–486 root, protecting, 369 secure, 264, 273 selecting, 295 sudo program and, 369–370 See also pseudo-random password path, adding directory to, 466 pattern matching and set statement, 391

O
Open Secure Shell (OpenSSH), 21, 254–256 operating system (OS) command syntax, output, and, 130–134 controlling case statement to pick, 437–438 exceptions file listing, 142 file system monitoring and, 128–130 /home filesystem, 142–143 monitor all OS listing, 134–141 See also specific operating systems

Index pdisk# all_defined_pdisks function, 501–503 all_varied_on_pdisks function, 503–505 cross-referencing to hdisk#, 520 list_of_disks function, 506–507 lsdev command and, 501–502 overview of, 495 translating hdisk# to, 496 percent (%) character, 103, 526 pg command, 392, 393, 399 PID (process ID), 263, 524 ping command, 37–38, 251, 319 pinging, automated hosts, with notification cron table entry and, 335 /etc/hosts file compared to list file, 333 functions, 329–331 listing, 324–328 logging capability, adding, 333–334 options for convenience, 321 overview of, 319 pager notification, 334–335 $PINGLIST variable length limit problem, 332–333 script in action listing, 331 syntax, 320 trap, creating, 323 “unknown host” and, 334 variables, defining, 321–323 pipe ampersand (|&), 232 piping to background, 231, 232 co-process to background, 31–32 file output to while loop, 57–58 to tee -a command, 352 plus (+) sign, 554 pop-up messages, sending to Windows to all users, 621–622 error notifications, 645 groups, adding, 623 to individual destinations, 623–627 log file and, 645 overview of, 43, 619 script for, 631–639 sending message, 629–630 testing user input, 627–629 positional parameters accessing value of $#, 45 overview of, 13–14 referring to, 601–602 pound (#) operator, 527 printers AIX classic print control commands, 404–408 AIX System V print control commands, 408–414 controlling case statement to pick OS, 437–438 exceptions capability and, 439 HP-UX print control commands, 414–417 keeping enabled, 38–39 Linux, controlling queuing and printing individually, 422–425 Linux print control commands, 417–422 log file and, 439 maintenance and, 439 printing and queuing all-in-one listing, 431–436 scheduling, 439 Solaris print control commands, 425–431 status information, 413–414 See also printing printf command, 41–42, 586–587 printing manual page, 465 page of passwords, 264, 271–272, 283–284, 294 See also printers; System V printing process ID (PID), 263, 524 processing file line by line cat $FILENAME|while line LINE method, 60–61 cat $FILENAME|while LINE=$(line) method, 61–62 cat $FILENAME|while LINE=`line` method, 59–60 cat $FILENAME|while read LINE method, 57–58, 77 command syntax, 53–54 file for testing timing of, 54–56 techniques for, 33 timing command substitution methods, 77–78 timing data for each method, 73–77 timing methods, 66–67 timing script, 67–72 while_line_LINE_Bottom method, 58–59, 76 while LINE=`line` from Bottom, 61

675

Free & Share & Open

676

Index processing file line by line (continued) while LINE=$(line) from Bottom method, 62 while line LINE with file descriptors method, 66 while LINE=$(line) with file descriptors method, 65–66 while LINE=’line’ with file descriptors method, 64–65 while read $FILENAME from Bottom method, 58 while read LINE with file descriptors method, 63–64, 76–77 process monitoring common uses of scripts, 248 end of process, 218–223 logging starts and stops, 223–228 modifications to scripts, 248–249 overview of, 215–216 startup loop, 216–218 timed execution, 228–230 .profile file, ownership of, 477 progress indicator countdown indicator, 91–96 creating, 35–36 overview of, 87 rotating line, 89–91 series of dots, 87–89, 95 ps auxw command, 179, 213–214 ps command, 23 ps -ef command, 216, 252, 257 pseudo-random number, creating filename, creating unique, 535–543 fixed-length numbers between 1 and user-defined maximum, 527–530 numbers between 0 and 32,767, 525–526 numbers between 1 and user-defined maximum, 526 overview of, 36, 523, 524 random number, description of, 523–524 shell script listing, 531–535 shell script overview, 530–531 software key creation and, 608 techniques for, 524 pseudo-random password array, loading, 280–281 building new, 282–283 creating, 264 functions, defining, 267–275 keyboard file, checking for, 280 listing, 284–294 loop list, building, 281–282 page of, printing, 264, 271–272, 283–284, 294 syntax, 264–266 testing and parsing command-line arguments, 275–279 trap, setting, 280 variables, defining, 266–267 pwdadm command, 385–389 pwd command, 466, 467, 471

Q querying system for name of shell script, 530 question mark (?), 28, 562 queuing. See printers

R random number, 523–524. See also pseudo-random number, creating read command, 53–54, 60 rebooting system, 337 redirecting standard error to standard output, 366 relative pathname, 470–471 remote command, running, 255–256 remote host, running commands on, 20–21 removing blank lines in file, 44 columns heading in command output, 45–46 repeated lines in file, 43–44 repeated line, removing from file, 43–44 resyncing, 313 return code, checking, 25–26 return command, 9 reverse video control commands, 392–393 highlighting text using, 391 Operations menu, 610–618 turning on and off, 154 root access, auditing, 476, 483–486 root access, restricted. See sudo (superuser do) program rsh command, 20–21

Index running commands on remote host, 20–21 pre, startup, and/or post events, 228–229, 249 printers, 38–39 remote command, 255–256 shell script, 3–4 silent, 28–29, 63 run queue, 194 rwall command, 24 pattern matching and, 391 removing blank lines from file and, 44 reverse video control and, 392 seed, 267, 524 select command, 42–43 sending pop-up messages to Windows. See pop-up messages, sending to Windows sendmail command, 34–35, 83–84, 330 Serial Storage Architecture (SSA) control functions, 501–507 disk identification, 495 error log, 520 executing commands, 520–521 identifying disks listing, 507–519 identifying disks listing explanation, 519–520 syntax, 496–497 usage and user feedback functions, 497–501 set -A command, 46–47, 265, 418 set statement, 391 sgid, 18 shell, 2 shell script comments and style in, 4–6 as interpreted, 2 running, 3–4 shift command, 14–15, 571, 607 show_all_instances_status function, 257 show_oratab_instances function, 256 silent running, 28–29, 63 simple_SQL_query function, 258 smbclient command, 619–621, 630 snapshot information AIX commands listing, 338–340 commands, selecting, 337–338 determining statistics to include, 367 functions, defining, 340–341, 352 listing, 341–351 listing explanation, 351–353 report output, 353–366 storing, 337 variables, defining, 352 SNMP (Simple Network Management Protocol), 86 SNMP trap, 85–86, 261

677

S
Samba downloading, 642–643 overview of, 24, 619 testing smbclient program, 643–644 See also smbclient command sanity check, 113–115 sar command, 179, 188–191, 197–203, 214 saving remote directory listing with FTP, 444–446 scale description of, 545 removing from scripts, 582 setting, 161, 165, 556 scheduling monitoring, 177 printers, 439 scope of variable, 14, 120 script command emailing audit logs, 493 logging user activity, 40, 478–479, 480–483 monitoring administration users, 489–492 options, 493–494 overview of, 475 repository for log files, 479–480 starting monitoring session, 479 syntax, 476–477 uses of, 477 searching for large file, 466–473 security monitoring user action, 475–476 pseudo-random numbers and, 543 See also passwords sed statement character substitution and removal and, 101, 102–103 highlighting text in file and, 38

Free & Share & Open

678

Index software dial-out modem, 84–85 license key, creating, 594–597, 608 See also Samba; sudo (superuser do) program Solaris df -k command output, 132 iostat command output, 187 ping command syntax, 320 print control commands, 425–431 sar command output, 190 swap command, 148–149 swap space monitor, 164–169 System V output, 409, 426 uptime command output, 183 vmstat command output, 192 special characters, escaping, 2 special parameters, 15–16 SQL+ database query, 258–259 sqlplus command, 257–258 SSA. See Serial Storage Architecture ssaidentify command, 496–497, 503 ssaxlate command, 315–316, 496, 504 ssh command, 254 stale disk partition, monitoring for automated execution, 316 disk subsystem commands, 298–299 event notification, 316–317 Logical Volume Manager (LVM), 298 LV level, stale PPs at, 299–304 overview of, 37, 297 PV level, stale PPs at, 304–307 SSA disks, 315–316 VG, LV, and PV monitoring with resync, 308–315 starting and stopping all printing and queuing, 409 startup event, 228–229, 249 stderr file descriptor, 54, 63 stdin file descriptor, 54, 63 stdout file descriptor, 54, 63 sticky bit, setting, 18 storing log file, 479, 493 string, testing, 47–50 style, 4–6 sub-shell, executing command in, 417 subtracting list of numbers, 556–561, 565 sudo (superuser do) program compiling, 371–377 configure command output, 371–375 configuring, 378–384 downloading, 370–371 /etc/sudoers file samples, 378–381, 381–384, 486–488 installing, 377 lecture message, 385 log file, 389–390 make command output, 375–377 monitoring administration users and, 492 need for, 369–370 Operations menu and, 618 overview of, 367, 369 script command and, 475, 476 using first time, 384–385 using in shell script, 385–389 suid, 18 SUN/Solaris. See Solaris su (switch user) command, 478, 492 swap command (Solaris), 148–149 swapinfo command (HP-UX), 147–148 swap space AIX paging monitor, 149–155 all-in-one paging and swap space monitor, 169–176 command syntax, 146–149 HP-UX swap space monitor, 155–160 Linux swap space monitor, 160–164 memory and, 145 options for, 176–177 paging space compared to, 145–146 Solaris swap space monitor, 164–169 symbol commands, 13 system information, gathering. See snapshot information system load monitoring detecting problems, 213 gathering data for plotting, 214 get_max function, 212–213 iostat command syntax, 186–188 iostat solution, 203–208 overview of, 179, 193–194 sar command syntax, 188–191 sar solution, 197–203 showing top CPU hogs, 213–214

Index uptime command syntax, 180–186 uptime solution, 194–197 vmstat command syntax, 191–193 vmstat solution, 208–212 System V printing AIX and, 408–414 commands for, 39 Linux and, 417–422 Solaris and, 429–431 timed execution for process monitoring co-process, 230–231, 245–246 getopts command, 218, 228, 229–230, 244–245, 246 in action listing, 248 listing, 232–244 overview of, 228 timeout, shell, 476, 486 time stamping process, 227 timing, at command and, 96 TOKEN variable, 570 top level down, 120 touch command, 446 tput command, 38, 389, 400–401 tput rmso command, 392 tput smso command, 154, 392 trap setting, 21, 280 SNMP, 85–86, 261 trap_exit function, 275 tr command, 24–25 trigger value, defining, 118 troubleshooting /etc/sudoers file, 494 proactive approach to, 403 See also snapshot information tty command, 223 twirl function, 499–500 typeset command number base conversion and, 41 overview of, 24–25 syntax, 585–586 variable length, setting, 529

679

T tail command, 45–46, 405 talk command, 24 tar format, 371 tee -a command, 352, 422 tee command, 223 testing binary numbers, 589–590 built-in tests, 26 character strings, 245 command input, 72–73 command-line arguments, 275–279 integers and floating-point numbers, 552–554 mail service, 82 null variable, 44, 115 numeric test comparison, 120 password file with variable exported, 455–456 password file with variable not exported, 456 response to system snapshot, 366 sanity check, 113–115 smbclient program, 643–644 string, 47–50 text strings, 24–25 timing of line by line processing, 54–56 user input, 627–629 text ASCII, 2 finding in large file, 391 highlighting in file, 38 uppercase or lowercase, 24–25 thrashing, 145–146 threshold variable, setting, 111 time-based script execution, 27–28 time command, 56, 67

U uname command, 128, 176 uniq command, 43–44, 622 Unix flavors. See AIX; HP-UX; Linux; Solaris until loop, catching delayed command output with, 32–33 until statement, 7–8 uptime command AIX system and, 180–181 field test solution, 184–186 HP-UX system and, 181–182 Linux system and, 182

Free & Share & Open

680

Index uptime command (continued) OS common denominator, 183–184 overview of, 179, 180 Solaris system and, 183 system load, measuring, 194–197 usage function, 274–275, 497–498 user capturing keystrokes of, 40, 475–476, 480–483 giving feedback to, 313 informing about monitoring, 493–494 logging activity of, 478–479 monitoring administration, 489–492 monitoring session, starting, 479 sending pop-up message to, 621–622 user information commands, 22–23 /usr/local/bin directory, 466 scope of, 14, 120 threshold, setting, 111 TOKEN, 570 verbose mode, 218, 222 Veritas filesystem, 495 viewing data assigned to variable, 13 visudo program, 378, 488 vmstat command, 179, 191–193, 208–212 volume group, 495

W wall command, 24 w command, 22 Web site for book functions on, 656–662 shell scripts on, 24, 647–656 Web sites Open Secure Shell code, 21 Samba, 642 sudo program, 370 which command, 640 while loop parsing file in, 53–54 progress indicator and, 88, 89 while statement, 7 who command, 22 wildcards, 28 Windows, sending pop-up messages to. See pop-up messages, sending to Windows Winpopup protocol, 620 write command, 24

V
/var, 493 variable COUNT, 570 double quotes (“) and, 16, 115, 392 global, 582 junk, 420 length, setting, 529 name of, and $ (dollar sign), 13, 151 null, testing for, 44, 115 overview of, 13 password, 456–463 RANDOM, 524–525 replacing hard-coded password with, 452–456