Very practical guide that does a fine job explaining recommender systems and how to set them up using Mahout. Sean notes that the book is well suited for beginners in recommender systems, and I have to agree. Unfortunately though, this book covers version 0.5 of Mahout. The current version, 0.9, has removed the SlopeOneRecommender and added many options for model-based recommendation. Coverage of the latter is lacking. However, for memory-based techniques, much of the content in this book is still viable and a good read for the beginner. Overall I would only suggest this book to beginners of recommender systems, and with a clear hint that Piero Giacomelli's Apache Mahout Cookbook covers version 0.8 and should be strongly considered.
EDIT: In hindsight I have to say that this book is the best I've found on Mahout. The author manages to address recommender issues in a way that the reader actually learns to understand the different implementation choices. Other works I've read focus on their own preferred way of implementing recommenders. Because those texts didn't bother to explain the alternatives they left me with more questions than answers. Not with this book. If you're starting with Mahout/recommenders, buy this book, you won't regret it.
I have a great interest in Big Data as part of my job and the understanding of machine learning and the applied algorithms that can be used to get more from your data is pretty cool - if you are a geek like me!