A required tome for anyone working with Solr. It is recent - actually, it is ahead of the curve - covering Solr 1.4 which is not even GA from apache yet (as of September 2009).
This is the only book for Solr. Literally -- nobody else has written one yet. Despite the lack of competition, the authors have done a good job putting some useful and new information to paper.
The book covers Solr and SolrJ - the embedded Java client API - and even provides some instruction on integration/embedding into your own Java app instead of using it as a stand-alone HTTP server. This capability exists but reference code and documentation is all but nil in the official docs. Performance tuning and replication are also covered. Generally, this book gives you what you need to make use fo the key (and some sideline) features of Solr so you can get it working for you.
And a big plus: no huge appendices of Javadoc that are useful only to increase page counts to make you feel you are getting 'value'. Really, who refers to Javadoc at the back of a book? I thank the authors and publishers for avoiding this temptation.
But the book does suffer from a problem inherent in tech publications: the assumption that the reader will start on page 1 and move forward. It tries to teach by creating one monolithic application that is spread throughout 300 pages. This is annoying if you start on chapter 8 as much context is lost. Also, these types of books spend too much time focused on the example application code and not enough time talking about the book topic. In the case of this book, the authors use a music database as their example application -- and spend many, may pages talking ancillary garbage about the music metadata, objects and the applications needed to download/use it. I don't care about how to use MusicBrainz. Really.
My free advice to publishers: develop your monolithic reference applications and post them to your website. Refer to them in your books, but use the pages otherwise wasted on music metadata trivia to show two or three alternate ways of using Solr, instead of the one that fits your fictional use case.
Solr has real potential but like most open source projects is sorely missing documentation and reference implementations. Like always, you could crawl the source to figure it out yourself ("What Would Stallman Do?") or buy this book. Frankly, my time is worth more than the book. I'd rather crawl through my own source.
While the authors get caught in the all-too-common approach of extending a single fictional use case throughout the entire book (hate that) - forcing too many pages to focus on the context of their use case and fewer pages on Solr - they do deliver enough of the goods to earn four stars. It would be five if they dropped the extraneous junk and used more pages for the core Solr product.
NB: I love Amazon but am going to point you to the publisher's website on this one. There you get the hardcopy book PLUS an immediately downloadable, password-free, copy-and-paste friendly eBook for about the cost of the paper version alone. I travel a lot and the eBook is amazingly helpful.
Amazon: add the eBook (not Kindle, I need this on my laptop).