I'm hoping you experts can help advise me, I've got this topology:
Live DB(Master) ----> Backup Server (Master and slave) ----> Test DB
Backup server is a slave of Live DB, and Test DB is a slave of Backup server
The replication from Live to backup server works as expected, the replication from backup to test is where I'm a little confused, if I create a test DB on backup, I can see it replicated to Test DB, but if I do the same on Live DB, it doesn't filter down to Test DB (but does to Backup).
Another thing I notice is that the binlog sizes on Backup are very small, only 150 Bytes, whereas they are close to 100M on Live, (binlog sizes set the same on live and backup)
Maybe this is by design? Maybe it's not possible to chain slaves together?
I goal of Test DB was (once working) to replicated from our Hosted Datacenter to our internal office, then use the internal slave to update our Dev / QA / Training DBs here (reducing bandwidth usage).
Here are what the related settings are on the backup server: (/etc/mysql/my.cnf)
log_bin = /srv/mysql-binlogs/mysql-bin.log expire_logs_days = 7 max_binlog_size = 100M binlog_cache_size = 16M binlog-format = MIXED
Specs: mysql 5.5 Debian Squeeze
Thanks, Your advice would be much appreciated