SCVMM VM location not always accurate

[Ctek]

Hi,

In one of our environments, we have a Backup Job that has a Hyper-V cluster as an object, so it takes all VMs in backup, 100-120 VM approximately.

This Hyper-V cluster object in the job is selected as part of a SCVMM managed server.

The issue we are facing is that some VMs are randomly failing to backup, and sometime works after retries. The error message is Unable to find Hyper-V hosts where VM 'GUID' is registered. We believe from the long troubleshooting we have done on our end, that somehow Veeam is not aware of the proper location (Which Hyper-V host) of the VM. When there is a 1 or 2 retries, it sometimes decides to work. We have disabled Dynamic Optimization so the VM are not moving automatically between hosts.

One of our sysadmins seems to experience a little less retries when we force a rescan of SCVMM in the Veeam console before executing the job. The culprit in this case can be Veeam, or SCVMM, but I was wondering if there is any registry tweaks to force Veeam to rescan more aggressively SCVMM maybe? Other suggestions?

This issue is bothering us since over 1 year and across different versions of Veeam (now 9.5 u3) and we had opened multiple tickets with Veeam to pinpoint the issue. The latest one being 02288367 I believe. It is not a high severity case, but it's a little PITA since it's a recurring issue.

Any suggestions? Ideas? Experience?

Thanks!

[nmdange]

I have only seen this issue when there is a problem with either VMM or Hyper-V. Typically if I refresh the VM inside VMM, it fails with an error. Usually live migrating the VM to another host and then refreshing again works. It has only happened a very small number of times to me though.

What version of VMM are you running? One of the new "features" in VMM 1801 is quicker/better host refreshes which might help with the problem. https://docs.microsoft.com/en-us/system ... c-vmm-1801

[Mike Resseler]

Hi Ctek,

nmdange is indeed right. This is mostly seen when there is a problem with one of the two, however, mostly I see the issue with SCVMM. I understand this is a PITA (I wouldn't even call it a little ). To be clear, none of those VMs have moved right? They have a successful backup, remain at the same host, and then out of the blue it states that it cannot be found anymore?

When you have this (and I am aware you need to be quick about it) does that VM have a yellow mark or red cross in VMM? Because that would also cause VMM to state that it doesn't know where the VM is. Normally this is seen because of VM movement (either live migration/ storage migration/ quick migration) and VMM is slow in refreshing. I haven't had the issue when it remains at the same host and nothing has happened with it.

Can you clarify?

[Ctek]

SCVMM 2012 R2 UR14

I understand what you guys are saying, in this case VMs are clean.

We are trying a pre-job script that does a Rescan-VBREntity -Allhosts, will let you know how it goes.

Thanks.