PostgreSQL 9 Administration Cookbook [Paperback]

As the book title suggests, it's a cookbook, but a cookbook that combines a question and answer style with a discussion style of writing. The tasks are neatly categorized into 12 chapters and each task smoothly builds on previous tasks discussed. It is still categorized in such a way that you can jump to a particular task you are currently having problems with without having read the other parts of the book.

Although it is titled PostgreSQL 9 -- it covers earlier versions as well.

This is one of those books I wish I had when training some of our customers or had early on. A lot of the questions - we are commonly asked or have stumbled on - like how to troubleshoot bad queries, how to tell what are my biggest tables, how to deal with data corruption etc, are all succinctly covered in this book.

Just to get a taste of what this book offers:

1. Chapter 1: First steps This is mostly a newbie chapter, that introduces you to PostgreSQL, guides you thru connecting to the database using commandline and PgAdmin. Some other examples of commonly used GUIs. Configuring access control, troubleshooting failed connections. It provides tips both for the Linux as well as the Windows user.
2. Chapter 2: Exploring the database starts to get into what I would call intermediate territory. It covers tasks such as determining where your database files are, how to determine disk space utilization for both whole database and individual tables. Getting quick estimate of number of rows for large tables where doing a count would be really slow. Using psql and the various system tables to determine object dependency.
3. Chapter 3: Configuration A good chapter not just for PostgreSQL users but I would say any database designer. I think this chapter holds useful nuggets for all levels of users. It covers basic considerations when planning your database and based on the needs how to configure the key PostgreSQL configuration settings by server, user, and database. It had tricks like how to determine which parameters are at non-default settings. I am ashamed to admit, that I learned a lot from this chapter. Stuff I really should have known before. My only lame excuse is that PostgeSQL has always worked so well, that I never had a need to delve into these nuances.
4. Chapter 4: Server Control - A very intermediate chapter slowly getting into more nuances. It covers connection pooling, viciously and not so viciously kicking users off, gracefully shutting down, using schemas to partition data. Again a lot of stuff here I didn't know so will have to reread when I find myself in these situations.
5. Chapter 5: Tables & Data - This chapter covers general good practices for namings tables and other objects, setting up keys, finding duplicate data, deleting duplicate data, . My favorite pasttime, techniques to generate test data. It concludes with loading data from flat files and csv files. Fundamentals of using pgloader utility created by Dmitri Fountaine.
6. Chapter 6: Security Covers new GRANT management in PostgreSQL 9.0, setting up users, auditing DDL and data changes, integrating with LDAP, using SSL, ensuring secure passwords, and encrypting data with pgcrypto.
7. Chapter 7: Database Administration - starts getting a bit into creating functions using plpgsql, plproxy. Running SQL scripts in parallel. Fundamentals of PostgreSQL DDL such as creating schemas, tablespaces and setting new configuration options for those introduced in 9.0. Moving objects to different schemas and tablespaces.
8. Chapter 8: Monitoring and Diagnosis - covers how to catch blocking queries, query monitoring. Using pgFouine for log processing and alerting you of issues. The thing I most enjoyed about this chapter were the cut and paste queries utilizing various PostgreSQL system tables to track down issues.
9. Chapter 9: Regular Maintenance - this topic is normally a very dry one,at least for me. It covers fundamentals you should know, like backup,recovery testing, reclaiming space, transaction wrap around (less of an issue with newer versions of PostgreSQL), but that puts a narcoleptic like me into a comatose state. I awoke when reading the topic of Carefully removing unwanted indexes. It shows you how to turn indexes off without removing them so you can see if they are really useful. I've really got to try this one out.
10. Chapter 10: Performance & Concurrency - My favorite chapter. Can't get enough of this stuff. Just read it. You'll learn a lot regardless your walk of life.
11. Chapter 11: Backup & Recovery - Covers using PostgreSQL specific tools such as pg_dump and other tools such as rman and rsynch for doing differential backups.
12. Chapter 12: Replication and Upgrades - It covers replication terminology, the new 9.0 streaming replication, hot standby as well as longtime favorites like Londiste, Slony-2.0 and load balancing with pgpool, and using pg_upgrade.

Disclaimer/note: I received a free review copy of this book from the publisher, in pdf format.

The PostgreSQL community has been waiting for a "cookbook", and I'm really excited that we finally have one. The PostgreSQL 9 Admin Cookbook contains many tips & techniques I'm going to put to immediate use. Like, pgloader and the ON_ERROR_STOP option to psql. This book also reminded me that there's still a lot I don't know about Postgres: for example, I've never worked with pg_controldata or the quote_ident() function. If you're a Pg admin (or wannabe), you should give this book at least a run-through, even if you think you already know everything.

Most of the "recipes" in this cookbook will stand on their own; some require material from previous or other referenced sections to make sense. I am a sucker for conversational style, and while the book gets off to a rough start, it does even out after a bit. Big blocks of SQL are formatted consistently in a style that, while it's not one I use myself, is easy to read.

The authors give a great explanation of why they prefer their filesystem set up a specific way, and how to do it (of course, it probably helps that I agree with them on this point); and good advice about schema & relation names. Specific problem-solving tools I found useful are: the list of steps to troubleshoot failed connections; specific things to do if a backend is hung, or a query is taking too long; and generating test data and taking random samples of real-world data. They also give warnings where something you do might cause application downtime.

Sadly, this book suffers from inadequate editing, dragging the rating down from a solid four stars. I often struggled against the grammar and organization to understand what the authors were trying to express. This frustrated me and slowed me down; a detriment in a technical manual. I also found that I couldn't rely on the (seemingly auto-generated) index to find what I need, although the TOC and chapter headings helped out a lot there.

Bottom line: would I buy a hard copy? At the listed pricepoint (US$45 at this time), probably not. However, it's a great first effort and I eagerly await the second edition.

This book does exactly what it advertises -- it tells you what to do and NOT to do if you want your PG application to fly.

I've read the pgsql-performance mailing list from Jan 2011 to now (Sep 5th), and you can find a lot of the stuff covered in the book, but it's not nearly as well organized. Save yourself the time and get more info by getting the book.