Linux System Administration Handbook

David A. Bandel

Issue #56, December 1998

Managers who don't actually administer systems, but make decisions regarding whether Linux is “appropriate”, or intermediate users who can use the better-fleshed-out chapters to complement other manuals they have, could find the book a good value.

  • Authors: Mark F. Komarinski and Cary Collett

  • Publisher: Prentice Hall Computer Books

  • E-mail: info@prenhall.com

  • URL: http://www.prenhall.com/

  • Price: $39.95 US

  • ISBN: 0136805965

  • Reviewer: David Bandel

The Linux System Administration Handbook was written by Mark F. Komarinski and Cary Collett. The authors claim to have nearly two decades of experience between them in system administration, most in Linux, though some in other UNIX systems. Reading the cover, I attempted to gain some insight into what lay in the pages ahead. The cover did promise much in the line of system administration tasks, security, hardware configuration, and much more. The book also included the gratuitous CD-ROM. (What Linux book doesn't these days?) The cover, however, gave no indication of the kind of audience it was aimed at, so I began reading, not exactly sure what to expect. The primary task I took on myself, then, for this review was to determine what audience the authors were aiming for.

The first few chapters cover boot-up and shut-down and a review of System V start-up, then dive into user administration and user shells. These chapters do not discuss Linux installation per se, but I am very familiar with Caldera's OpenLinux Lite, the distribution included with the book, and it really doesn't require much in the way of installation instructions, at least not for anyone vaguely familiar with Linux installation. Besides, all of Caldera's pertinent documentation is included in the Appendices.

As I continued reading I was impressed with some of the jewels of wisdom and little-known (or not well-documented) facts about Linux that only seasoned administrators would know. Unfortunately, this was often offset by short, terse explanations, short chapters and one- or two-sentence summaries for those basic chapters toward the front of the book.

Going a little further and getting into the coverage of nitty-gritty network administration, the chapters grow longer. Like an administrator who's finally been given his favorite project to work on, the chapters suddenly take on more detail and more life. The authors go into great detail explaining some of the lesser-understood and lesser-used of the well-known services, the kinds of problems you can expect, and how to configure, troubleshoot and maintain them. In fact, most of the chapters where they went into this kind of detail are quite well-done. Much of the information presented showed the authors do indeed know a good deal about those programs and services to which they have had personal exposure.

Continuing on, they discuss some of the important issues for decision makers wanting to know about applications for Linux. Here, they took on the daunting task of trying to do justice to all the applications beginning to show up for Linux, from open-source software to commercial native Linux applications to those which can be adapted to Linux. They did a creditable job and warned the reader they would discuss only those programs with which they had some familiarity. However, I was still surprised they didn't do a bit more homework for the reader, including a few more applications which they don't use, but are available.

For example, in the section on databases, they mention that Oracle, while not supported on Linux, can use the SCO binaries with iBCS. While they don't mention it, the same is also true of Informix. Native Linux applications such as Adabas and YARD (both from Germany) were not mentioned. I find this even more curious since Adabas is sold by Caldera, and YARD rivals Informix in its ability to do nested outer joins and other complex SQL queries. YARD is also ANSI SQL 92 and SQL3 compliant, something most open-source Linux databases can't begin to boast about. The authors talk about distributions later on and mention WGS, which positions the Flagship database as its premier Linux product, but this isn't mentioned in the database section.

While I was hoping that this would be a good book to help beginners discover Linux, mostly due to the exceptionally easy-to-install OpenLinux distribution, it is hardly that. Some omissions and skimpy coverage of basics would lead me to conclude that this is not a good book for a novice administrator. For example, in the rather short chapter on “Common Features”, where the authors discuss setting environment variables under the bash shell, they don't mention the export command, its usage or implications. This kind of oversight could have novices wondering why subshells or programs invoked by the shell haven't inherited a particular environment variable.

On the other hand, I also can't recommend this book to experienced system administrators on the strengths of its detail in the lesser of the well-known services, even though these sections are well done. While the book would make an excellent addition to a library lacking the details of network news (NNTP) or other services, it isn't justified because of the light or non-existent treatment given to other areas.

I can, however, recommend it to managers who may have noticed their system administrators have begun to use Linux on their network or are contemplating allowing this to occur. The book does a good job of introducing Linux in a way that would give managers a good feeling about this oft-called “Renegade OS” being put to work in their companies.

On a scale of one to five, I would have to rate this book a solid three. Its apparent lack of focus, its terse coverage of some important areas, and redundant coverage of some network issues among the Networking and the Internet Connectivity chapters make me conclude this first edition doesn't deserve a higher rating.

For those who think I may be a little harsh, my initial impression was much lower. While I may be tolerant of a few clichés or awkwardly worded phrases, the authors trounced solidly on a pet peeve of mine throughout the first few chapters by misusing the phrase “try and” when they really wanted you to “try to” do something.

However, I hope the authors will soon begin work on the second edition, because I can see the framework for a good Linux system administration handbook. Some other things I hope the authors will consider is either changing the distribution to Red Hat, which is what they talk about in many examples, or changing their examples and discussions in the book to reflect the distribution. For those who don't know, Caldera OpenLinux, while using the RPM system, is not a Red Hat distribution, but is based on the German LST distribution. I would also like to see more discussion of the File Hierarchy System (FHS), partitioning schemes and disk recovery with emphasis on fsck, and other pertinent commands. Hopefully, they'll also close some of the really gaping holes, like forgetting to even mention one of the major Linux distributions, Debian, not to mention all the foreign distributions, such as S.u.S.E.

Overall recommendation for this book: look twice before you buy. Managers who don't actually administer systems, but make decisions regarding whether Linux is “appropriate”, or intermediate users who can use the better-fleshed-out chapters to complement other manuals they have, could find the book a good value. Otherwise, I'd hold out to see if the next edition is a more solid investment.

David Bandel is a Computer Network Consultant specializing in Linux, but he begrudgingly works with Windows and those “real” UNIX boxes like DEC 5000s and Suns. When he's not working, he can be found hacking his own system or enjoying the view of Seattle from 2,500 feet up in an airplane. He welcomes your comments, criticisms, witticisms and will be happy to further obfuscate the issue. He can be reached via e-mail at dbandel@ix.netcom.com.