Am 16.02.2011 12:33, schrieb Carl:
> The max_allowed_packet setting is the same on both.
the question is how large the setting is
we have 200M on all machines
> I have tried restarting the slave... didn't work
after replication errors you should every time
* stop the slave
* "hot" rsync the dadadir on the master
* stop the master
* "cold" rsync to get last changes
* remove binlog files in datadir und backup
* start the master
* rsync the backup to the slave
* start the slave
* restart replication
this way the downtime of the master is very short because
the second rsync must only copy changes since the first run
Attachment: [application/pgp-signature] OpenPGP digital signature signature.asc
