The latest replicated snapshot is what is used to create the zero copy clones on the remote side. Judging from the error message, it seems as though it can't find the snapshot, and therefore doesn't know what to create the clone from. However, I understand from your post above that you are indeed replicating. Perhaps a quick call in to Nimble support is in order?
Thanks, after a bit of digging with support we noticed that checking the "replicate recent changes" option allows it to work. It is a bug. If not checked it is still looking for the snap that SRM would create rather then the most recent available. It exists in 3.4 and 3.5 and is currently a priority one bug.
Now I am curious to see if it is isolated to the test fail over, so I am loading up a test vm in its own group and attempt a real fail over for the 1 vm without replicating changes. (To simulate an actual disaster) Just have to wait for it to replicate.
For the sake of completeness: I tested an actual failover on a test vm, shutting down the sra server on production side to prevent a replication. Worked as expected. So the bug is in the testing mode it will not zero copy clone the most recent snap, only one that is replicated by the test. Where in the actual recovery it just claims the volume and works fine.
Ok I have been rolling out my first deployment of SRM 6.1. Things seemed to be going great.
Ran through all the checkboxes for SRM. Including add array pair.
All relevant volume collections are replicating, have several copies at the remote site.
So I hit the test button, and it quickly fails on step 4. For every server and every vol collection:
Error - Failed to create snapshots of replica devices. Failed to create snapshot of replica consistency group XXXX. SRA command 'testFailoverStart' failed for consistency group XXXX. The volume collection that you specified cannot find the latest replicated snapshot
I log in to the remote nimble and where I would expect to see a ton of zero copy clones, I do in fact see nothing new. So it is never actually creating the clones. And my google-fu is proving less than effective.
Did I miss something?