I am trying to setup a DC-DC setup in openstack victoria, I have 2 numbers of all in one setup (controller, compute) with shared mysql and rabbitmq cluster and am using ceph image replication and configuring cinder replication on top of it.

So when the 1st node goes down then i will perform a 'cinder failover-host node1'  and then do a nova-evacuate to failover to the 2nd DC setup and once the node 1 comes up i will use cinder failback and live migration from node2 to node1.

**i am facing some minor issues in this setup which I will ask once I know this is the right approach to the DC-DC concept.

Can you please shed some light on this being the right approach or not and if not then how can i improve it?

With regards

Swogat pradhan