This book is 400-some pages and includes a CD-ROM of utilities and such that the author uses for examples. The first thing to mention is that this book is geared to a very specific sort of cluster - one that is intended to load balance practically any type of connection request that would start a job. This could be batch jobs or it could be transaction processing, such as with a database.

The first part of the book deals with Cluster Resources. In some ways, it's a review of such basic things as compiling the kernel and using netfilter. At first I thought it was a waste of time, but I ended up appreciating his review and pointing out a few things that I hadn't considered, that could be important further on.

The second part deals with High Availability. The author describes the relatively simple two systems with failover capability which can be used as is, sort of like a mini-cluster, or for a redundant load balancer which would be required for anything larger.

The third part deals with building the cluster, proper. There are several methods, and the pros and cons of each are discussed.

The fourth part deals with Maintenance and Monitoring, so you are not left with a system that, once built, will fall apart on you.

I liked the detail and references provided, for what was provided. I was a bit disappointed that the back end (data storage) was not addressed much at all. The author basically described using NFS or a NAS box, but really gave little detail. Mind you, the book was already, as mentioned, over 400 pages.

I find that the book is completely useful for what I'm intending to use it for. I'll just have to dig a bit deeper for my own ideas on the data storage.