we have the following setup currently:
-DD2500+NBU master server in main site+clients
-DD640+NBU Media server at DR site
-data is written to VTL and storage pools/units
- replication is done over DD, no SLP.
we wanted to test restore in DR using the Media server in DR and the data present in the DD640, that seemed to be impossible due to not having a catalog in DR,
now we are planning to make the media server in DR a standby master server, import the catalog and try to restore from VTL and storage units,
do you think the restore is feasible in this scenario?
how can the standby master server detect the replicated storage units?
I don't want to be very critical but the proposed approach raises quite a few questions of techical nature, for example how you are planning to implement a standby master from a media, is it the same domain or different? are you going to shut down your production master when running from the standby?
You actually can restore in DR provided the DR media server can connect to the master in production to query the catalog, any other scenario requires DR for the master server and replication of just backup images is not enough - you should also replicate the catalog or the master itself using either a compatible replication technology or building a cluster. Alternatively you can take a catalog backup and restore it to a temp VM/server with the same name as your production master. This new server should be either isolated in the network or you primary master should be shut down while you do DR restores to avoid naming conflicts.
Then it's a no-brainer. You restore the catalog to a stand-by machine, "export" relevant tape in DD console (check EMC guides how to do it) and Bob's your uncle. You just need to make sure that the DD in DR is zone to the server you will be using as the media server and your NBU has that virtual robot configured and run robot inventory to update the EMM.
All these steps are pretty well covered in Dell EMC guides.
i will use the latest full catalog image present into the DR destination on the current master server and recover the catalog into the new standby master server.
after completing the catalog recovery, i will test restore from VTL replicated Mtree and DDBoost replicated Mtree as oer below:
- for VTL replicated MTree, I will zone the VTL to the new master server through DD console and try to restore from the replicated tapes.
- for DDBoost replicated Mtree, I will enable ddboost on the replica MTree using the following command ddboost storage-unit modify <MTree-Name> user <ddboost-username> so the replicated MTree becomes assigned STU and then i will try to restore the images from this STU.