Hyper-V 2016 checkpoint and SAN snapshot behavior

[StephanF]

Hi,

we are just starting to migrate to Hyper-V 2016 and realized that checkpoint and SAN snapshot behavior is quite different to prior versions. Veeam with Hyper-V 2012R2 always took a SAN snapshot using the hardware VSS provider in on-host and off-host mode. With Hyper-V 2016 Veeam is only using production checkpoints in on-host mode and production checkpoints + SAN snapshots in off-host mode.

Former user guide versions stated for Hyper-V 2016 that in off-host mode a Hyper-V production checkpoint is created, then a hardware SAN snapshot is taken and after this the production checkpoint is released immediately. This made perfect sence to me but it wasn't working this way. We opened a support case (# 02277027) and now user guide has been changed to reflect the current behavior. The production checkpoint stays during the whole VM backup in parallel to the SAN snapshot and is released after the backup finished.

My questions:
Can anybody explain this behavior?
Doen't this mean we need double snapshot space (once for the Hyper-V checkpoint and once for the SAN snapshot)?
Why is the Hyper-V checkpoint needed after creating the SAN snapshot?

Thanks
Stephan

[DGrinev]

Hi Stephan,

After a brief discussion with the dev team, they told me that behavior is by design due to changes that have been implemented in Hyper-V 2016 framework.
Released checkpoint triggers the guest OS for log truncation and post backup processes, that's why the checkpoint should live before the end of the backup.

Doen't this mean we need double snapshot space (once for the Hyper-V checkpoint and once for the SAN snapshot)?

That's correct. Thanks!

[StephanF]

Thank you for the information.