O'Reilly's Linux System Administration - for MCPs?

by Tom Adelstein

The marketing people at O'Reilly seem to get right to the point when they describe their books. On the catalog page of Linux System Administration you will find this quote:

The ingredients for this book had been scattered throughout mailing lists, forums, and discussion groups, as well as books, periodicals, and the experiences of colleagues. Everything is now in one handy guide. In the course of their research, the authors also solved many problems whose solutions were completely undocumented. They now pass their lessons on to you.

As the primary author, I can explain the inspiration and some of the issues involved. When we started the project, I didn't expect it to take two years of concentrated effort to get the result. But, like the copy states the answers to many OSS application issues required much research. If I wanted to see this book come to fruition, then I had to persist and that meant late nights and long days searching for hints on the Internet as the starting point.

So, Why MCPs?

Aside from my life as an Open Source practitioner, I'm also a MCP going through recertification on the Server 2003 track. Integrating Linux application servers into domains and forests seemed easy to me, but getting the applications installed on Linux proved challenging far too often. As the paragraph above says, many solutions were undocumented. So, we had to document them.

I had an inspiration on how to put the book together, but when writing a proposal you need prove a market exists. So, I went looking for the market. In an interview with Mike Webber at Spidertools I recognized the need for a definitive Linux administrator's book unlike any I saw in the market. When I interviewed Mike, he called from his headquarters in Trout Creek, Montana. His business model involves putting a customer on a box and training him or her one specific technology at a time such as building a Sendmail server, etc.

Then I ran across Falko Timme, a German fellow who started publishing comprehensive, step-by-step guides to building Linux servers. I built a Debian server from scratch using his "Perfect Setup". His approach of building a application servers one at a time proved a smart way to put material together.

Some of the work at the Open Source Software Lab at Microsoft involves research to help Microsoft practitioners work with Linux- or perhaps stated differently - help them integrate technologies. I understand the problems the Lab faces as I faced them eight years ago. As a Solution Provider, I needed to integrate a Linux box running an Apache Server into an ecommerce solution. The first step in that process required learning how to use Linux, which provided quite a challenge.

Now, imagine this type of scenario. An enterprise system admin at contoso.com gets a call from the CIO who wants a load balanced High-Availability Apache Cluster on Novell SLES 10 running Heartbeat setup for a demonstration in a week. OK, the admin says, I'll just head on over to a search engine and run a few searches. As bits and pieces of information comes forward, panic starts to emerge. That's how I felt eight years ago.

Today, he can sit down and open a chapter in a book and build the server in a couple of days. That's why I wrote Linux System Administration. You won't find much discussion about Linux this and Linux that, but you will learn to build a server, get the components together to create a load balanced Apache cluster and other applications box by box.

I'm recertifying on Server 2003 and some of the best material I found sits on the Technet site in Step-by-Step guides. So, this isn't a new concept, I just never saw the approach taken at the system admin level in the OSS community until I ran into Falko's web site. And, by the way, I don't discriminate when I say system administrator or system engineer. It's all technology to me and I want to learn as much as I can as quickly as I can.

Does this Benefit the Open Source Community?


It does in several ways and O'Reilly's marketing people once again articulate the points:

Linux is now a standard corporate platform ... and there is a definite shortage of talented administrators. Linux System Administration is ideal as an introduction to Linux for Unix veterans, MCSEs, and mainframe administrators, and as an advanced (and refresher) guide for existing Linux administrators who will want to jump into the middle of the book.

In the early days of Linux almost everyone contributed code. The community started off small and grew to around 30,000 users by the release date of Windows 95. By 1998, the Linux community reached around 2 million and only a small percentage of those folks contributed code or could administer even a web server.

I saw this book as a way to go beyond the standard Linux Bibles and power user orientation. I knew many people who wanted to deploy stand alone servers running Apache, DNS, Mail, Blogs, etc. I also saw experienced administrators wanting to learn how to put together high performance clusters, virtual machines and servers running multiple silos on the same box or blades and to incorporate single sign-on directory services into their infrastructures.

Technology moves quickly and if you have to find answers from mailing lists, forums, and discussion groups the time to find out how to do something takes longer than doing it. From my perspective, it takes significant resources to put together the documentation to allow people to quickly deploy the solutions they need. Those resources should come from major players such as the consortium at OSDL and the troopers at the Open Source Software Lab.

I came to recognize the massive proliferation of NT infrastructures when I started up my first Linux company. The first task required someone to build a UNIX client for Exchange. With pockets of Solaris and Linux users in companies like Ericsson, Boeing and Intel to mention a few, a need existed to join the two communities. We call that interoperability.

I see more acceptance today on both sides of aisle - a bi-partisan effort so to speak in the fields of technology. So, as the days pass, I expect everyone to begin working together to create a world that works for everyone. Only time will tell if it comes together in pristine form.