Live Migration between two WS2012 R2 Clusters with SCVMM 2012 R2 creates multiple objects on Cluster

Hi,

I'm seeing an issue when migrating VM's between two 2012 R2 Hyper-V Clusters, using VMM 2012 R2, that have their Storage provided by a 4 Node Scale Out File Server Cluster that the two clusters share.

A migration between the two clusters is successful and the VM is operation but I'm left with two roles added to the cluster the VM has moved to instead of the expected one.

For example: Say I have a VM that was created on cluster A with SCVMM, resulting in a name of : "SCVMM Test-01 Resources"

I then do a live migration to Cluster B which has access to the same storage and then I end up with two new roles instead of one.

"SCVMM abw-app-fl-01 Resources" and "abw-app-fl-01"

The "SCVMM abw-app-fl-01 Resources" is left in an unknown state and "abw-app-fl-01" is operational.

I can safely delete "SCVMM abw-app-fl-01 Resources" and everything still works but it looks like something is failing during the process.

Has anyone else seen this?

I'll probably have one of my guys open a support ticket in the new year but was wondering if anyone else is seeing this.

Kind regards,

Jas :)

December 19th, 2013 12:25am

You're not alone. I've been testing migrations between a standalone 2012 R2 host and a 2012 R2 cluster using VMM 2012 R2 and it's also creating the duplicate cluster entries.

Free Windows Admin Tool Kit Click here and download it now
December 28th, 2013 11:39am

Cheers for the reply.

Doesn't appear to be breaking anything, just making a mess of the labelling for the clusters, DPM and SCVMM.

Glad I'm not the only one as we have a mirrored configuration for our storage here and I thought it was related to that.

All my team are off until the 3rd but I may get one of them to open a ticket then unless it starts to become more than an irritating "feature".

If you have issues with this can you drop a further reply so I can keep an eye on it as I have about 200 production VM's that have been moved seeing this problem.

Unfortunately I couldn't halt the process without seeing issues.

Jas :)

December 29th, 2013 4:42pm

I'll update the post if I see it become an issue. I only have ~100 VMs to move once my testing is complete, so it'll be a bit of an annoyance to clean up.

--James

Free Windows Admin Tool Kit Click here and download it now
December 30th, 2013 7:08pm

I have seen this myself as well. 

The common thing in these situations, is that the VMs were created outside of VMM in the first place, and not from VMM directly. This means that the SCVMM suffix prior to the VM name, is missing, and the migration process will recreate with new suffix. 

It is true - it does not break anything, but certainly it is a mess, to have a copy of the VM that is offline.

Please note that I have not been able to test different workarounds here, but it could be an idea to rename the cluster resource role prior to the migration (not the VM name, but the name of the cluster resource = VM name in failover cluster manager).

-kn

January 2nd, 2014 9:38pm

In my case, the VMs were created on standalone hosts via VMM 2012 R2, then migrated to the cluster using VMM 2012 R2. So, VMM was involved the entire time.

I'd understand the issue if it was done outside VMM, but that isn't the case for me.

Free Windows Admin Tool Kit Click here and download it now
January 2nd, 2014 9:49pm

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics