[Pacemaker] Mysql Replication Problem with HA processes

[Andrew Beekhof]

On Fri, Oct 7, 2011 at 7:34 PM, raki <krakesh872 at gmail.com> wrote:
> Hi Andrew
>
> Hi Andrew
> We had installed Heartbeat/Pacemaker on two Unix(Centos 5.0.2 Machines)
> The RPM's we had used for installing Heartbeat ad Pacemaker are
> heartbeat-3.0.3-2.3.el5.i386.rpm
> pacemaker-1.0.9.1-1.15.el5.i386.rpm
> Please find the crm(cluster resource manager) order  and co-location we had
> used.
> colocation Httpd-with-Mysql inf: HttpdVIP MS_Mysql:Master
> colocation Httpd-with-ip inf: HttpdVIP Httpd
> colocation Mysql-with-Tomcat inf: Tomcat1 MS_Mysql:Master
> colocation Tomcat-with-HttpdVIP inf: Tomcat1 HttpdVIP
> order Httdp-after-HttpdVIP inf: HttpdVIP Httpd
> order Httdp-after-tomcat1 inf: Httpd Tomcat1
> order MYSQL-after-HttpdVIP inf: MS_Mysql HttpdVIP
> we had two nodes running HA processes (cluster with two nodes)
> In on of our test scenario we tried to restart the node where Mysql-Master
> is running, based on the above configuration the Ha processes restarts and
> other node Mysql processes takes over the Master responsibility
> And we saw the Mysql replication stops working and in the Slave status we
> found this error duplicate entry for the key values.
> Please help us regarding this.
>  Error Description  'Duplicate entry '2083' for key 1' on query. Default
> database: 'MSF_DB'.

I'm not really skilled in the workings of mysql replication.  Perhaps
the author of the RA can comment?

> I ca also provide what ever information required.
> waiting for you response on this.
> Rakesh