Discard Changes in Recovery Site

[smudger]

Quick question and open to your greater knowledge...

When the VM(s) have been failed to the recovery site, we have certain cases where the changes made (post business testing) are not needed so can be discarded. Obviously re-protecting the VM will mirror those changes back to the primary site which is not intended or required.

Our implementation is SRM 5, using basic non funky SAN based replication (3PAR) and ESXi 4.1, although could bump to v5.

Could VMware based snapshots, or cloning be used ? Once the VM is failed over and testing has finished, the clone is then failed back and destroyed/cleaned down.

Outside of SRM this could be fairly straight forward ie. stop replication, split the disks, the remote site becomes read/write etc, however I need to achieve the same using SRM.

Has anyone else had this requirement, I'd be really interested to know?

[Vandalay]

Sounds like your use-case was to discard data after testing the plan.  Why not run the RP in Test mode which accomplishes what you're looking for out of the box?

Otherwise, you could maybe create a script as a custom step in the recovery plan to snapshot the VMs as they are recovered.  Then remove the snapshots before failing back to the original site.  You could be limited in how long you can stay in the failed over state, however, due to the usual issues with leaving snapshots on for too long.

[TheITHollow]

I agree with Vandalay about using the "Test" Recovery Plan.  Then when you're done, you'll just blow away the failed over VMs and go back to your production network.

If this isn't an option, you could try to do a failover, and then instead of doing a re-protect, build another recovery plan that only fails over the VMs that you need to fail back to the production site.  Then you can power on the VMs that you failed over to the DR site, but decided not to fail back.  I'd do some testing first though

I hope this helps.