Hi,

please, refer to emc216579 "How to conduct an application test on destination side without stopping replication sessions" .

In this case, you don't need to stop the replication sessions, and use a crash consistent image of the production file systems on the destination side.

Let me know if this fits your needs.

Gustavo Barreto.

Thanks for the information. Actually we are doing that currently for testing purposes and what I was looking more for like an actual DR testing where we failover the filesystems and disgard the changes. Is there anyway we can do that using  "reverse" option ?

Hi,

I don't think you can do a reverse in this case.

You would need to do the following :

1) switchover the replication sessions;

2) work on the destination side as you need;

3) delete the replication sessions;

4) remount the original source file systems as RW, and set the VDM's state to loaded ( this will bring the CIFS servers online );

5) manually undo the changes on your DNS environment ( if needed - The DDNS should take care of that );

6) remount the original destination side file systems as RO, and set the VDM's state to mounted;

7) recreate the replication sessions.

But, I strongly recommend you to try first on a lab environment, so you can prepare your settings on both sides, and become more comfortable with the procedures. Create a test CIFS server on a test VDM, and a test file systems with some test shares...

Keep in mind that you will have a production downtime from step 1 to step 5.

Gustavo Barreto.

We don’t have a test NAS to need to do it on the production box but as you said with test filesystem and test cifs server. we are also not using vdm’s. so will test it ahead. information helps a lot. thanks once again…

Another quick question. if we delete the replication session then does it to a full copy again once we re-create a session ? thanks

Hi Manumari,

I have similar test coming up in my DC  as  you described in this post  where i have to discard  all the changes made to original destination  site ( when it is acting as source upon failover)   and  return back to start state from where we started the test  ..Were you able to perform the  test in  you DC ..  could  you pls  post the steps  ..

Thanks in Advance..

Yes. I was able to test it successfully. Following were the steps I performed:

1. Took a checkpoint of the production Celerra before the change just in case to rollback

2. Failover over the shares using “switchover” flag which would be a clean failover

3. In switched over state, Prod would be Read-only state and DR would be Read Write

4. Once testing is complete, Manually change Prod shares to Read write & DR shares to read-only

5. restart the replication back to DR

Any questions do let me know.

Manumari

I have a issue, when using the "switchover" method. The switch over to DR goes through fine, The DNS updates with the new ip, and the shares are mounted on the switched over VDM.

The problem is that I cannot write to the shares. Although the new source (DR) file system shows that it is in fact R/W/.

I have posted in this forum a procedure that allows you to bring up resrouces at the DR site.  This is to use ONLY if the two locations are isolated and cannot see each other during the DR test.

I'd be happy to review that with anyone.

                     source      destination   replicating
  Initially          RW    -->     RO             Y
  Switchover     RO    -->     RW             N     

To switch back...

Open a Putty session to the destination:
     server_mount server_2 -option ro FS_NAME 

Open a Putty session to the source:
     server_mount server_2 -option rw FS_NAME
     nas_replicate -start rep_NAME -overwrite_destination