2012/11/12 Machiel Richards - Gmail <machiel.richards@stripped>
> Hi Manuel
> Please take careful note of what I have stated in the original
> 1. the fact that the slaves say "seconds behind master = 0" does
> not neccesarily mean that repliication is working as I have found this a
> lot of times where it shows 0 then have proven that replication was not
> working (I.e. data not being processed on the slave). The 0 seconds simply
> means that it has read all the data but does not mean that it has been
> processed or up to date.
Agreed. What I meant is that replication is not broken (slaves are able to
connect to the master). If they're not replicating data, that's another
> 2. the Exec_Master_Log_Pos and Master_Log_File are in fact out of
> sync with the master (i.e. on the master the file and log position are way
> ahead of the slaves by about 9 hours).
Have you tried to take a look at the master's binlog to see what happened
at the position where the slaves got out of sync? Just to make sure it's a
> 3. We have tested replication by creating a test table in the
> database with some data and it was not replicated to the slaves.
I am assuming that Replicate_Do_DB is empty in both slaves, isn't it?
Have you tried to start replication again at the same position and file in
one of the slaves to force them to force them to rotate their relay logs?
That's what is usually done when the relay bin gets corrupted (although you
get an error message in the show slave status output)