Why are the best technical books also the thinnest? To make the case, I highly recommend the MySQL 5.1 Cluster Certification Study Guide In rack units, it’s a 1U; it fits nicely into my laptop bag; and if you’re considering implementing MySQL Cluster, it can save you a world of time.
Of course, if you get the book, you should consider the certification itself. There’s a legitimate debate about the usefulness of certification exams, but the MySQL 5.1 Cluster certification is a little more important than others for a couple of reasons:
1. In case you haven’t heard it by now, Cluster isn’t always the best fit. In many cases, an active/passive failover setup is a much more flexible and cost-effective approach. It’s not always a clear-cut decision. If you’re willing to make application changes and if you understand the internal workings of MySQL Cluster, it can be an amazing solution.
To see if cluster is a good fit, I recommend the following guides:
http://www.mysql.com/why-mysql/white-papers/mysql_cluster_eval_guide.php
http://www.mysql.com/why-mysql/white-papers/mysql_ha_solutions.php
2. There aren’t many consultants and DBA’s that have actually installed and used MySQL Cluster. There’s still a lot of misconceptions. The primary misconception is: “I know MySQL, I’m sure that I can quickly set it up and support a MySQL Cluster.” Outside of MySQL professional services group, there’s limited expertise in the marketplace. If someone has a MySQL Certification, at least you know that they’re tall enough for this ride.
I thought the most important part of the book was Chapter 7 regarding Indexes in NDB tables. Unless specifically suppressed, a hash and ordered (T-tree) index will be created for primary keys. The hash index is stored in the IndexMemory pool. The ordered index is stored in the DataMemory pool.
If you want to create a unique index, that’s a bit more complicated and it takes about a page to explain. Unique indexes have performance implications, which provides even more reason that someone should really understand MySQL internals before implementing a cluster.
To summarize, I’m hesitant to say this in public, but frankly; the study guide is a good read. It won’t make the Oprah book club any time soon, but it has some good discussions about hash indexing vs. t-tree indexes, database isolation levels, ACID properties and high availability concepts.
[sidenote: I usually rip technical books in half to achieve 1U. Usually, you can safely throw the first half away.]
