Thank you for your answer. On master we got "incorrect kay file" error
on the same table in error log. I cannot tell what caused that. We tried Innodb on that
table, but applications didn't work properly.
--- On Tue, 12/27/11, shl_256 <shl_256@stripped> wrote:
> From: shl_256 <shl_256@stripped>
> Subject: RE: Slave got " 'Invalid error code' (126)"
> To: "'Johan De Meersman'" <vegivamp@stripped>, database100@stripped
> Cc: "'Replication'" <replication@stripped>
> Date: Tuesday, December 27, 2011, 10:45 AM
> One of the likely causes (barring
> unusual replication setup) for differing data, is
> nondeterministic queries, like "insert into table1 select *
> from table2 limit 10": because there is no order by in the
> select, it is possible for the select to return a different
> set of records on the master and the slave. There's plenty
> more examples like that.
> Maybe, it could not happened in InnoDB,
> for it default ordered by clustered index. Before this SQL
> execute, the data between master and slave are the same and
> -----Original Message-----
> From: replication-return-2288-shl_256=163.com@stripped
> On Behalf Of Johan De Meersman
> Sent: Thursday, November 17, 2011 3:34 PM
> To: database100@stripped
> Cc: Replication
> Subject: Re: Slave got " 'Invalid error code' (126)"
> ----- Original Message -----
> > From: "Database System" <database100@stripped>
> > [Error] Slave: Query caused differnt errors on master
> and slave.
> > Error on master: 'Invalid error code' (126), Error on
> slave: 'no
> > error' (0)
> Exactly what it says: a query yielded different results on
> the master and the slave. As your databases are supposed to
> be identical, that is seen as a critical error and
> replication is halted.
> First of all, is this a normal replication setup? No
> variations in engines, no excluded tables, et cetera ?
> Use the Maatkit tools (mk-table-checksum --count
> <masterhost> <slavehost> | mk-checksum-filter)
> to verify if there are no differences in data between the
> master and the slave. It's very likely that you'll find
> some, or the replication would have kept working.
> One of the likely causes (barring unusual replication
> setup) for differing data, is nondeterministic queries, like
> "insert into table1 select * from table2 limit 10": because
> there is no order by in the select, it is possible for the
> select to return a different set of records on the master
> and the slave. There's plenty more examples like that.
> Another possibility, of course, is someone messing about in
> the slave's data :-)
> Bier met grenadyn
> Is als mosterd by den wyn
> Sy die't drinkt, is eene kwezel
> Hy die't drinkt, is ras een ezel
> MySQL Replication Mailing List
> For list archives: http://lists.mysql.com/replication
> To unsubscribe: http://lists.mysql.com/replication?unsub=1