From: Robert Citek Date: February 14 2013 10:55pm Subject: Re: slave replication with lots of 'duplicate entry' errors List-Archive: http://lists.mysql.com/mysql/228964 Message-Id: MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 On Thu, Feb 14, 2013 at 4:54 PM, Manuel Arostegui wrote: > 2013/2/13 Robert Citek >> >> On Wed, Feb 13, 2013 at 8:59 AM, Robert Citek >> wrote: >> > Any other possibilities? Do other scenarios become likely if there >> > are two or more tables? >> > >> > Of those, which are the most likely? >> >> [from off-list responder]: >> > Other possibility: The replication is reading from master not from the >> > point when the dump was done, but some time before and is fetching insert >> > statements which are already in the dump. >> >> To prevent that I used the coordinates in the dump file included with >> --master-data=2. Could the coordinates still be off? > > > Hello, > > Are you sure nothing is getting inserted directly into the slave? Is it in > read only mode? > If you're starting replication using the values provided by --master-data=2 > (which should be something like): > > -- Position to start replication or point-in-time recovery from > > -- CHANGE MASTER TO MASTER_LOG_FILE='mysql-bin.000974', > MASTER_LOG_POS=240814775; > > And if you're using the right IP, there's no reason to have duplicate > entries unless someone is writing directly into the slave. > > Manuel. According to the client, nothing is writing to the slave and everything is being logged at the master. I have not had the opportunity to independently verified any of this, yet. I do know that the slave is not in read-only mode, but rather "we promise not to write to it" mode. At the moment, I am trying to come up with plausible explanations for the observations. Thanks for your feedback. Regards, - Robert