Am 09.01.2014 17:54, schrieb Sergey Spiridonov: > Unsafe statement written to the binary log using statement format since > BINLOG_FORMAT = STATEMENT. UPDATE IGNORE is unsafe because the order in > which rows are updated determines which (if any) rows are ignored. This or > der cannot be predicted and may differ on master and the slave. > Statement: UPDATE IGNORE dbmail_mailboxes SET seq=seq+1 WHERE mailbox_idnr=2 > > If I put option binlog-format=MIXED into my.cnf, warning goes away. > > Is it a good idea to use mixed binlog format in this case? Are there any > hidden drawbacks?
there are hidden drawbacks *not* doing so "statement" is pure danger and "mixed" is also not really sane if you want to avoid all sorts of troubles you should even use binlog-format = ROW i changed to this 2 years ago because i had enough of breaking replication randomly on different server (independent of dbmail)
signature.asc
Description: OpenPGP digital signature
_______________________________________________ DBmail mailing list DBmail@dbmail.org http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail
