From: Shawn Green (Shawn.GreenSun.COM)
Date: Sun May 03 2009 - 12:38:49 CDT

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Dominik Klein wrote:
> Hi.
>
> I have a question regarding mysql replication and mysqldump.
>
> I have a master (A). All my clients insert/update/delete only to this
> master. Then I have a Slave (B). This slave only replicates the master.
> There are no other processes changing/inserting data into the Slave. The
> slave also logs binlog so I could replicate from that server as well.
>
> Now I want a chained Slave ( like A -> B -> C , C being the chained slave).
>
> So my idea is: stop replication on B so no changes during dump, dump its
> master status, mysqldump all databases. Then load the dump on C and
> configure slave on C according to the master status from B.
>
> I did that and end up in hundreds of duplicate key errors. How can that
> be? What should I do now? Do I need to wait for some settling after I
> have "stop slave" on B and before starting the actual mysqldump?
>
> Mysql Version is 5.0.51b on A and B, 5.0.77 on C, operating system is linux.
>

If you did this:

(on B)
STOP SLAVE;
SHOW MASTER STATUS;
<dump all data>

(on C)
<restore all data>
CHANGE MASTER TO ... <binary log coordinates from B>
START SLAVE;

Then this should have worked.

If C started replicating from the first binary log of B, then this would
very easily explain the mass of duplicate key problems.

Double check your replication coordinates.
--
Shawn Green, MySQL Senior Support Engineer
Sun Microsystems, Inc.
Office: Blountville, TN

--
MySQL General Mailing List
For list archives: http://lists.mysql.com/mysql

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]