Hi!
On Oct 24, Baron Schwartz wrote:
> In this particular case, the program would have retried the statement that
> deadlocked:
>
> INSERT into MyISAM temp table
> SELECT from InnoDB table join MyISAM temp table join MyISAM temp table
>
> >>We have been having persistent issues with statements causing
> >>deadlock, yet being written to the binlog anyway, thus replicated
> >>through to the slave, where of course there is no deadlock. The
> >>error message is "Query caused different errors on master and slave.
> >>Error on master: 'Deadlock found when trying to get lock; try
> >>restarting transaction' (1213), Error on slave: 'no error' (0)"
You're mixing transactional and non-transactional tables in one
transaction. Such a transaction is always written to binlog, even when
it's rolled back. It is assumed that changes to transactional tables
were rolled back, changes to non-transactional tables remained, and the
same will happen on the slave - so changes to non-transactional tables
will be replicated correctly. On rollback you must see a warning "Some
non-transactional changed tables couldn't be rolled back".
Of course in your case the rollback doesn't happen on slaves :(
Try to ignore this error with --slave-skip-errors=1213
Regards,
Sergei
--
__ ___ ___ ____ __
/ |/ /_ __/ __/ __ \/ / Sergei Golubchik <serg@stripped>
/ /|_/ / // /\ \/ /_/ / /__ MySQL AB, Senior Software Developer
/_/ /_/\_, /___/\___\_\___/ Kerpen, Germany
<___/ www.mysql.com
