The Ceph volume driver now includes RBD replication, which provides replication capabilities at the cluster level. This feature allows you to set a secondary Ceph cluster as a replication device; replicated volumes are then mirrored to this device. During failover, all replicated volumes are set to 'primary', and all new requests for those volumes will be redirected to the replication device.
To enable this feature, use the parameter replication_device to specify a cluster that the Ceph back end should mirror to. This feature requires both primary and secondary Ceph clusters to have RBD mirroring set up between them. For more information, see http://docs.ceph.com/docs/master/rbd/rbd-mirroring/.
At present, RBD replication does not feature a failback mechanism. In addition, the freeze option does not work as described, and replicated volumes are not automatically attached/detached to the same instance during failover.