Hi All :) I am creating an environment based on about 15 hardware nodes. 6 of them will be for mysql databases. They will be divided into three pairs (3x2 nodes). Nodes in every pair will be configured with replication.
I done similar configuration about an year ago. I used than: - percona 5.5 with standard asynchronous replication; - Multi Master Replication Manager for MySQL (http://mysql-mmm.org/) for automatic assignment/failover of read and write IP addresses and checking the status of replication (mmm provides status scripts); - Pacemaker to create a cluster in which scripts for Multi Master Replication Manager were used. My question is: I heard that there are other options than standard asynchronous replication which sometimes was problematic (for example there was one slave thread only in percona 5.5 and there was a possibility that slave node will fall behind master node in high load situations). I also am thinking which MySQL fork is the best option if I plan to use replication. What can you propose based on your experience? BR, Rafal.