DPM 2012 R2 keeps failing on hyper v backup of 2 x 2012R2 servers

[dapaulio]

The VSS application writer or the VSS provider is in a bad state. Either it was already in a bad state or it entered a bad state during the current operation. (ID 30111 Details: VssError:The specified object was not found.
(0x80042308))

 

Hi All,

This has been bugging me. I have 2 servers that show this error when they backup. if I remove them from the protection group and add them again they data is transferred and show successful, however every subsequent backup then fails "replica inconsistent" and every attempt at a retry shows the above error with 0mb transferred.

 

VSS writers all show stable,

dependent services are in a started state

 

I have several 2012r2 servers in the same protection group that don't fail, just these 2 in particular. to the best of my knowledge nothing has changed with the server or dpm.

 

I feel like it is staring me in the face

Any suggestions?

[sbrade47]

Had something similar to this when trying to back up system state on 2012r2 hyper-v VMs, the VMs were part of a failover cluster so in the end I changed it to back up the VM vhdx instead under the cluster name. Then just protected any worthwhile data separately avoiding system state.

 

I know it struggles with system state on Hyper-V VMs with or without the clustering. Used to fail all the time.

[dapaulio]

Thanks with your help i had a light bulb moment when you mentioned cluster. I looked at the cluster and I don’t know what specifically fixed it as I removed the cluster notes on the vm, (vaguely recall this being the culprit for some anomalies a few years back) then deleted the servers out of the protection group, refreshed the cluster discovery within dpm then added the server again.

I then noticed that it duplicated each server in the protection group.

On initial creation one of each duplicated succeeded and the other failed. Removed the two that failed and left the two that were successful. Forced a recovery point and looks positive this time. I will find out after this weekend when the schedule goes through whether it has worked. Thanks again for your help

[sbrade47]

No problem. Glad you've got it sorted! I battled with this for ages before I realised I was doing it wrong and that as it is a VM I don't need system state anyway.

 

Fingers crossed it will be stable and working tomorrow morning. I had a few occasions where I thought I'd cracked it and was badly mistaken though!

 

After I posted I realised I'd explained the cluster node bit badly too. Just for the benefit of any others reading this I backed up VMs by expanding the cluster name and choosing the VMS from within. In the management console they will show as ServerName.ClusterName.DomainName.

Edited October 22, 2018 by sbrade47