Not much to report other than the fact that I am pouring over HowTo’s for building and deploying MySQL clusters and chewing up my insides with anxiety over the added wrinkle of tossing them up in to Amazon’s EC2. After some heady and heavy reading on the subject I’m feeling a little bit like Gab’s up there. Hopefully, in the next couple of days I can get something working and maybe, just maybe, I’ll have a quick and dirty How-To of my own posted if all is a resounding success–you can be certain if it isn’t there’ll be a post about it!
You’re going to have a tough time with MySQL on EC2 due to the lack of permanent disk storage. Mark Atwood got a great deal of attention just for getting it to read from S3.
If you’re not complete set on EC2 try TheGridlayer.
Yeah, immutable storage is one of my main concerns and the solution seems to be replicating to at least one data node that resides in a standard configuration but beyond that there are some nagging doubts that plague the concept.
The biggest is security, since most most clusters are built behind a firewall and reside on a privet network are4 a little easier to secure but EC2 doesn’t support private subnets (at the moment) and setting up a VPN will kill performance. What to do?
The other concern is the cost. While it’s pretty cheap to deploy a server or two launching a load balancer, cluster manager, two data nodes, and a SQL server could get a little pricey.
Thanks for the tip, though, I will definitely check out TheGridLayer.