[Pacemaker] Multi-site configuration for PGSQL RA
Takatoshi MATSUO
matsuo.tak at gmail.com
Sat Nov 2 03:37:24 UTC 2013
Hi Richard
2013/11/1 Richard Colley <rcolley at cardaccess.com.au>:
> I have just setup a postgres 9.3 cluster using the pgsql ra. It's all
> working now (despite much hair pulling to start with :).
>
> I would like to extend this to multiple sites.
>
> I've decided not to try and have a single cluster across the WAN. Instead,
> I have 2 independent clusters.
>
> I thought to use a ticket to transition a cluster between the "active" &
> "passive" state. Creating the ticket & creating a ticket constraint was
> easy enough.
>
> But now I need to do something with each site's cluster when this change
> occurs. In particular, I need to make sure all postgres instances in the
> passive site are in slave mode and replicate from the vip of the promoted
> site. The active site will just use the standard pgsql RA behaviour.
>
> Has anyone done anything like this before?
This ?
https://github.com/t-matsuo/resource-agents/wiki/Resource-agent-for-postgresql-9.2-for-disaster(split)-site
You can use rep_mode="slave".
> If not, I might try and update the pgsql ra with this capability. Any
> advice? I note that the mysql ra has this capability, so I might refer to
> it's implementation.
>
> Thanks,
> Richard
>
>
> _______________________________________________
> Pacemaker mailing list: Pacemaker at oss.clusterlabs.org
> http://oss.clusterlabs.org/mailman/listinfo/pacemaker
>
> Project Home: http://www.clusterlabs.org
> Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
> Bugs: http://bugs.clusterlabs.org
>
Thanks,
Takatoshi MATSUO
More information about the Pacemaker
mailing list