RabbitMQ DR configuration w/ Pacemaker

[Zach Anderson]

I'm attempting to setup RabbitMQ with the following configuration to allow for a disaster recovery site configuration.



I have the local clusters configured and I'm attempting to add them as a pacemaker resource:

pcs resource create rabbitmq ocf:rabbitmq:rabbitmq-server-ha

In this configuration though, the resource can only run on 1 of the nodes at a time so my local cluster is ineffective.  Does anyone have any guidance on how I can set this up so the local cluster is still usable but with the ability to fail over to the DR site?

Thanks!

[Zach Anderson]

So I found the clone options and that seems to be working for what I needed.

pcs resource clone rabbitmq

RabbitMQ then runs on both "granted" nodes at the same time as desired.  One further question I have is how to configure the automatic fail-over.  The base booth behavior seems to only revoke/grant the ticket if the "granted' nodes are unreachable.  Can you configure it so that if the resource group is in error (not able to start on any node) that it will revoke/grant to the other site?