other than what Gleb says,
you can avoid the error using 'insert ignore into' or better yet
'replace into'
for every insert statement
AESYS S.p.A. [Enzo Arlati] wrote:
>I'm trying to use teh parameter sql_slave_skip_counter at run-time to
>restore slave replication.
>When a slave replication broke due some errors in code, my be a duplicate
>key, the only working way to restore the replica where to delete the
>existing record which conflicts whith the ones inserted by the replication
>process.
>So if I have a duplicate key 30020 ,I have to remove the record with the id
>30020 and the replication can reinsert it's copy of record with id = 30020.
>This should be difficult to automate so I try another way using the global
>variable sql_slave_skip_counter.
>
>I try to skip 5 records using a statemente like this:
>set global sql_slave_skip_counter = 5;
>and then restart the slavre
>start slave;
>
>but the problem is the same and the command 'show slave status' report a
>skip_counter filed equals 0.
>
>Where I am wrong , some can help me ?
>Regards, Enzo
>
>
>
