MySQL database replication
Contents
Description
We use mysql replication to mirror mysql databases from a single master mysqld server to a series of slave servers on back end nodes. This process is asynchronous and asymmetrical - back end nodes are read-only.
Configuring the Replication Master server
Two options must be set to enable replication. First, binary logging must be enabled. Second, every mysqld server in the pool needs a unique ID.
These parameters should be set in the my.cnf file (/etc/my.cnf on WormBase servers).
[mysqld] log-bin=mysql-bin server-id=1
brie3
(the primary WormBase development machine) acts as the master server, with server-id = 1
.
Configuring the Replication Slave servers
The only required configuration for replication slaves is the server-id:
[mysqld] server-id=2
Binary logging is not required for slaves but can be useful for data recovery and more complicated architectures. I have it off on our read-only slaves for performance.
See WormBase_Infrastructure and Hardware catalog for server IDs.