P2000 snapshot using hardware vss provider give error CAPI_ERROR_NOT_A_MASTER_VOLUME

[Christophe Van Lierde]

We have a P2000 G3 SAN attached to a Hyper-v 2012 server ( SAS connected ). Latest Capi proxy and VSS Provider drivers are installed. Latest SAN firmware is installed.

We create a backup job (Veeam) configured to trigger the hardware vss provider. The job includes several virtual machines. During each VM backup, veeam triggers the snapshot, backup the VM and then releases the snapshot.

This only works though for the first VM : veeam triggers the snapshot, a snappool is created on the SAN, a snapshot is created for the volume, the VM is backed up, the snapshot and snappool are deleted.

When Veeam tries to backup the second VM, it tries to trigger a vss snapshot again, but an error occurs and in the application log this is logged and the backup fails :

capiCreateSnapshots(): Capi Error 0x000000AC (Error: CAPI_ERROR_NOT_A_MASTER_VOLUME) on operation reply.

All snapshots followed after this, give the same error. Only when the event '

The VSS service is shutting down due to idle timeout." is logged on the Windows 2012 hyper-v server (after some idle time) then the first snapshot will succeed again, all following wil fail again etc ...

Anyone seen this happening?

Christophe

[Dave Minto]

Hi

I have exactly the same issue with a 2 node Windows 2008 R2 SP1 Hyper-V cluster and a P2000 G3 SAS SAN.

How did you resolve the issue?

Regards

Dave

[Aurock]

It sounds like your SAN is deleting the snap pool and converting the master volume to a standard volume when the last snapshot is deleted.  There is a registry key you can modify to change this behavior.  

From pages 24-25 of the HP StorageWorks MSA System VDS and VSS Hardware Providers Installation Guide: 

( http://h20000.www2.hp.com/bc/docs/support/SupportManual/c02520792/c02520792.pdf )

Snap pools are automatically deleted
In order to help manage limited disk space and snapshot licenses, the VSS Hardware Provider will, by
default, automatically convert master volumes to standard volumes when the last snapshot for the volume is
removed, and automatically delete snap pools when no snapshots refer to them.

However, this mode of operation may cause problems in environments where some manual management
of snap pools is desired, or in SAN environments where multiple hosts are using VSS to create and delete
snapshots concurrently on the same array.

To avoid these issues, the VSS Hardware Provider can be configured so that it does not attempt to
automatically create or delete snap pools or convert standard volumes to master volumes and vice versa.
Instead, the administrator must perform these steps manually.

Solution
To disable automatic conversion of standard volumes and automatic creation and deletion of snap pools,
enter the following command at the Windows command prompt to change the provider's
StandardVolumesAreSupported parameter value to 0:

reg add "HKLM\SYSTEM\CurrentControlSet\Services\HP P2000 MSA2000 VSS Hardware Provider\Parameters" /v StandardVolumesAreSupported /f /t REG_DWORD /d 0

To restore the default behavior, change the value of StandardVolumesAreSupported back to 2:
reg add "HKLM\SYSTEM\CurrentControlSet\Services\HP P2000 MSA2000 VSS
Hardware Provider\Parameters" /v StandardVolumesAreSupported /f /t REG_DWORD
/d 2
After changing the StandardVolumesAreSupported value, the VSS server must be stopped if it is currently
running:
net stop VSS
net stop "HP P2000 MSA2000 VSS Hardware Provider"
Alternately, reboot the server to ensure that all the VSS components have been restarted.