I am running a MySQL master DB (v5.5) on an Amazon Linux AWS server. This is linked to two slave servers by standard MySQL replication (with a binary log).
The setup works perfectly, right up until I have to reboot the master server -- at which point it seems to flush or reset the binary log (always resetting it to the same number) but fails to pass this on to the slaves, who become confused and stop replicating.
It's easy enough to reset things manually but there must be a config I am missing to stop this happening.
I have tried shutting down manually using mysqladmin rather than relying on the SysV shutdown scripts, but that made no difference. The MySQL server is started using the default SysV init script.
The my.ini file contains these lines related to replication:
log-bin=/var/run/mysqld/logs/mysql-bin.log
binlog-do-db=litir
server_id=1