I am new to using RBR and am having problems recutting a database that uses it.
Previously, with statement based replication, we would just stop another slave database at a know slave position, take a snapshot of the disk pack and then mount this snapshot up on the machine I am recutting and copy all the data over. We’ve never had any problems using this method.
I have trying to do this on a large-enough-to-be-annoying (700GB) database, but when I have copied all the data over and put it back into replication I get a whole bunch of errors like so:
Last_Error: Could not execute Update_rows event on table dbJbossQueue.JBM_MSG_REF; Can’t find record in ‘JBM_MSG_REF’, Error_code: 1032; handler error HA_ERR_KEY_NOT_FOUND; the event’s master log vmbu04-bin.002645, end_log_pos 32580193
They are not all updates, the errors happen with Delete’s as well. I have also done the cut twice to try and eliminate me messing up.
I am starting to wonder whether it is possible to recut a database that uses RBR using this method, or whether I need to use mysqldump or similar instead?
p.s. I am using Percona-Server-5.5.13-rel20.4-138.Linux.x86_64 and XtraDB