Intermittent VM recreation failure, stuck in NeedsSysprep

[ConnorO1]

We have six RDSH VMs on a Hyper-V cluster that are automatically recreated each morning at 5-minute intervals.

Intermittently, one VM fails to complete recreation. The affected VM is successfully deleted, cloned, powered on and the Parallels Guest Agent connects, but RASPrep does not complete.

In multiple confirmed failures on different VMs, vdiagent.log shows the same sequence:

- NeedsSysprep → RASprepInProgress
- Rejected to update agent state
- NeedsSysprep → RASprepInProgress
- Rejected to update agent state
- NeedsSysprep → OK
- Rejected to update agent state


Successful VMs instead progress:

- CloningInProgress → RASprepInProgress → OK
- Preparation completed successfully
- Error code [0]

This does not appear VM-specific as different VMs are affected.

We are running Parallels RAS 21.2.1.1 build 27310, installed since mid-August, and are upgrading to 21.2.2.1 build 27432.

Has anyone seen RASPrep become stuck in NeedsSysprep with RASprepInProgress and OK state updates being rejected?

I will not attach the log file itself due to size and business specific naming.

Any help would be appreciated

 

[Thierry FRACHE]

Hi Connor,

This looks more like a Hyper‑V VDI Provider state-notification race than a VM-specific RASPrep issue. The repeated rejected to update agent state messages suggest that RAS is receiving stale or conflicting state updates while the VM is being prepared.

Please verify that all connection brokers, providers, agents are on the same build.

21.2.1.1 release notes mention a critical fix for VDI provider agent responsiveness with Hyper‑V clusters and fixes for RDSH recreation failures. I would test the issue after upgrading all components, preferably to the latest 21.2.x build, then retry with the VM recreations serialized or with a longer interval.

Hope this helps !

Thierry