oracle

This book is somewhat mis-titled as a STATSPACK book. Instead the book is a superb overall view of Oracle tuning using STATSPACK.

The concepts are presented clearly and elegantly, and the supplemental scripts are very useful. This is a great book to add to your collection that you will actually use.

the bad:
1. no mention of Windows anywhere. only covers UNIX, no mention of this in the blurbs
2. Oracle 8i (and even 8.05!) take up more pages than 9i. It's rarely clear where 9i info starts. An obvious copy+paste job from previous editions. And this despite the title of the book.
3. the author spews scripts everywhere. there should be a table of contents for the scripts. it's practically impossible to figure out what scripts there are and what they do. Possibly really lazy or inexperienced DBAs would be interested in all these scripts but most dbas will probably live without them.
4. author promises to provide "extensions" to statspack tables for trend analysis and he makes it sound like he will present a clever way to store snapshot data between reboots the way Capacity Planner does, but it turns out all that all he demonstrates is how to create tables for data collected by UNIX operating system utilities. It is a very misleading description.

the good:

1. decent explanations of basic tuning concepts. But again this is copy+pasted straight from a tuning book. I bought this book for info about statspack not tuning 101!
2. decent diagrams of statspack tables but nothing that can't be found on metalink.

I returned this book. It's shovel-ware, plain and simple. The author worked very hard to bloat the size up to 600+ pages to rationalize a $40+ price tag for the publisher. How can he take himself seriously? Or i wonder if the author actually believes this is a well-organized book to be proud of...in which case his mind must be like a tangle of knotted strings.

In my opinion this book is for junior to senior dba's. It covers basic, advanced and even expert DBA topics. The books pays for itself over and over agin.