i had configured a MySQl Replication Server with 2 DB Servers.Now we are planning to implement a failover Scenario.If Master fails The slave should act as a Master and we have a addition server.It will be going to act as slve.i had mentioned the points below..


Replication
(Master Server)----->----(Slave Server)-----------(Additional Server)
DBS 1 DBS 2 DBS 3

1.Replication is made between Master (DBS 1) and Slave (DBS 2) (successfully done)

2.If Master (DBS 1) failed Slave (DBS 2) should act as a Master and Additional Server (DBS 3) needs to be Slave

3.Whenever Orginal Master (DBS 1) became online the current Master ie ..Slave (DBS 2) should go back to Slave position and all lost data will be replicated to Master DBS 1.

how can i meet this scenario..Any Suggestion will be appreciated..

Thanks
Ajayan

Hello,

Have a look at this chapter at dev.mysql.com. It has everything you need to configure your situation and the steps to follow when the master fails.

Kind regards,

Eric

1 members found this post helpful.

Thanks My Friend..
It was a Nice article and now i got a good idea.I will try to work out this.

Since i Working with Critical data,is any option to automate this failover Scenario otherwise i have to keep an eye to Change Master-Slave Settings on each server.

Thanks,
Ajayan

Hello,

In order to monitor the status of your servers, it might be handier to write your own script using Bash or another scripting language and have it send an email or other notification when a node is not replicating correctly or when the master is down. If you have a monitoring server like Nagios you can also include it in there.

Kind regards,

Eric

What about Using HA/DRBD..Will it make any Sense?.

Hi,

Of course that would make sense. Have a look at HeartBeat in combination with Pacemaker. I don't have experience with DRBD so I cannot elaborate on that option, but I'm sure there will be someone on LQ who'll step in.

Kind regards,

Eric