Failed to create VM recovery checkpoint

[Frosty]

Okay, so this is not-strictly-speaking a Veeam issue, seems to be a Hyper-V / VM / something issue ... but I thought the good folk here might have seen this before and be able to put my investigation in the right track.

Started getting this error on our (soon to be retired) on-prem Exchange hybrid server yesterday when running our daily backups:
"Failed to create VM recovery checkpoint (mode: Veeam application-aware processing) ... etc"

Ran the job a couple of times. Failed every time. Rebooted the Exchange server VM. Backup job kept failing. Tried taking a manual snapshot of the VM using Hyper-V Manager on the host server. This also failed. Logged on to Exchange VM and ran Powershell. vssadmin list writers returned nothing. Checked Windows services. Volume Shadow Copy service was running, but I restarted it just in case. vssadmin list writers still returned nothing. Microsoft Software Shadow Copy Provider was not running (startup type=Manual). Started it. vssadmin list writers started returning a long list of VSS writers. All were Stable and no problems observed. Ran Veeam backup job and it worked.

Any suggestions as to what is going on with this VM, and/or why the Checkpoint creation process is getting broken?

As best I can tell this only became a problem yesterday (all previous backup jobs were working) and it just so happens that I installed the latest 2025-11 Windows CU for Server 2019 yesterday. Suspicious much. Has anyone else noticed backup issues since this NOV Windows Update?

EDIT -- I should add that this VM was previously running on vSphere and was migrated to Hyper-V using Veeam Instant Recovery -- so maybe something in VSS broke somewhere during that transition?

[david.domask]

Hi Frosty,

Not known issue at first blush. I checked quickly with colleagues from our Support team and nothing immediate came to mind either.

May be worth checking through System / Application event logs and seeing if the Microsoft Software Shadow Copy Provider service had issues and was set to manual after multiple failures or something like that.