We have a use case of going live in production with Hbase on a Mapr Cluster .MaprDB is not an option here at the moment and would be considered as a migration at a later stage.We are aware that with hbase on a mapr cluster, there is no possibility of have volumes to segregate your data.All the data going in the volume called /hbase . In such a scenario we are having discussions of what can be the best way to have a DR using Mapr's Volume Mirroring.
We think that we can setup a mirror for the volume /hbase into another volume called /hbase-mirror on the remote cluster.This would mean they actual volume /hbase on the DR cluster remains in touched(will not become read only) and if needed hbase can operate on it.During a DR Switch we can point /hbase-mirror to be /hbase and operate.During the failback to primary , it will once again be changd to /hbase-mirror and the delta data can be pulled back to the prod cluster.
Is there any other better way of doing this? Or does the above approach might have some problems?