We’re doing some research into how to scale our MySQL database from the current Master and Two Slaves set up, to a scalable cluster, ideally without going down the master to master replication route MySQL offer. In doing so, Wade stumbled across the “Spider” storage engine, but information seems sorely lacking and the home page does very little to help.
From what we’ve read so far, the Spider Storage Engine is a drop-in replacement for InnoDB. The engine transparently handles both sharding of storage and distributed query processing across a pool of MySQL nodes. Upon creation, you define the column it should look at, and then the range to distribute to each server, and everything else is handled for you. It also appears to be Drizzle compatible. However, apart from that, we know very little else.
I’m hoping that someone reading this will have used Spider in the past and be able to give their opinions. If you have used it, please drop a comment below with your experiences. We want to know about both positives and negatives, any benchmarks, any glaring pitfalls. We’ll likely post about our experiences if we end up testing it, but until then, whatever knowledge people have would be very much appreciated.
