Debugging and tuning of Linux, though not a romantic subject, is a necessary one for the health of any well-run free software biased network. A slight change here and a nudge there in the background can make the difference between user discomfort and a quiet Sunday afternoon for the poor old administrator. Linux Debugging and Performance Tuning as written by Steve Best is an excellent zoom in on this complex and detailed subject area.

The book’s cover

One significant issue I have with Linux, and yes of course it is a positive one of success, is that luckily for the poor power user there is a large collection of tools for debugging and performance tuning. Which set of optimized tools should any tidy administrator or power user choose? My first impression of the content of this book was that the author knows what he is doing and is leading the reader to the gold nuggets and the range and depth of the book accurately reflects the technical markets reality.

Steve Best has led the Journal File System (JFS) for Linux project. This fact implies that he has experience right down into the kernel of the kernel. His book of 476 pages and 14 chapters contains a dense set of descriptions of the tools you need to get a grip on the debugging and performance tuning requirements of your situation. He starts by simply explaining how to profile using such tools as time and date. However, that was only a well thought out tactic to bring the user gently into the warm folds of the subject area. Moving through code coverage, debugging, memory management, system information and speeding towards dynamic probes, the book reflects the clear wealth of knowledge required to get the majority of issues to a positive resolution. Best mentions many tools to the expected detail so that you will have a clearly developed understanding of which of the current range of tools best serves you under any given condition. I personally liked the description of the “Magic Key Sequence” to get a Back trace (page 174). Such details would be hard to discover under pressure. I would not have found this nugget by googling alone.

A well thought out tactic to bring the user gently into the warm folds of the subject area

Who’s this book for?

Quoting from the introduction of the book, and I have no reason to disagree, this book is intended for the person who is developing or supporting Linux applications or even the Kernel. I would go further and say that the power user that likes to run the make command raw may also enjoy this work. I even dare to say that this is a development team book and would sit nicely on the team room bookshelf next to the coffee pot and the almost dead potted plant.

Relevance to free software

Linux has many features that are significantly better than the proprietary competitor’s attempts. The kernel has a reputation for stability and predictability. Keeping this hard won reputation during the fight for Enterprise dominance is supported by maintenance of real instances by system administrators. Further, device driver makers play a crucial role in the wide spread support of the hardware real users use. If a device driver is buggy then the kernel may panic. Wide range device driver support requires wide range debugging. Steve Bests book helps quickly to get the relevant parties up to speed on the tools of the trade. Thus, this is a potentially important book focused on a crucial target audience and delivered at the correct moment.

If a device driver is buggy then the kernel may panic

Pros

Linux Troubleshooting being a book with a fast pace that veritably speeds through a large toolset, this book would sit snuggly on most Linux developer’s tables. The density of information allows for a broad overview. The title exactly describes the content. Thus, a well thought out book

Cons

Being a practical book about tools that help with performance or debugging it is unlikely that the average desktop user would gain value. Further, if you have one tool in mind that you particularly wish to explore in detail then you may need to look elsewhere. However, once you knew a product defined in the book exists you could then zoom in on the technical details via documents found on the internet.

In short