SMHV jobs fail after live migration of Guests

[rdenyer001]

We have been  doing some testing with  our 4 node Windows 2008 R2  Hyper-V cluster  and SnapManager for Hyper-V.

We have some test Windows 2003 and Windows 2008 guests some with and some with out pass though LUNS.

Last week  we had finally applied all the relivant hot fixes as per NetApp Hyper-V best practice and NetApp Technical support, and were able to get 100% sucessful jobs SMHV jobs of both types of guests with and without pass through LUNs.

To continue  testing our  IT manager  did a live migration to move the guests  around  on the cluster, as you would if you were going to take some hosts off line  for maintanence.

Our SMHV jobs started failing with the following VSS  errors

Backup failed with VSS Retryable error.
(9)] Writer Microsoft Hyper-V VSS Writer involved in backup or restore encountered a retryable error. Writer returned failure code 0x800423f3. Writer state is 5.
For more information, look at the Hyper-V-VMMS event log in the Windows Event Viewer under "Applications and Services Logs->Microsoft->Windows".

Any ideas on what  we have missed  ?

We were expecteing that  SMHV  would not be affected  by  live migrations ??????

Regards,

Richard

[rdenyer001]

Further update

After a day of  un-installing and checking all configuration options for SnapManaager for Hyper-V and SnapDrive  we have found that its is only the  VMs with pass through LUNs that fail.

[ophir]

Hi ,

I have the same issues with live migration and the server doesnt have Pass through disks ,

I have lots of Virtual machine that failed .

Backup task failed to succeed. Error: Vss Requestor - Backup Components failed.Writer Microsoft Hyper-V VSS Writer involved in backup or restore encountered a retryable error. Writer returned failure code 0x800423f3. Writer state is 5.
For more information, look at the Hyper-V-VMMS event log in the Windows Event Viewer under "Applications and Services Logs->Microsoft->Windows".

Any Idea ?

Thanks

BTW do you have the all list of the Hotfixes that install ?

Ophir

[rdenyer001]

Hi,

Needed to make sure  that Hotfixes were installed

In particular  we needed to take this in to account  as a lot of our guests VMs had pass through LUNs

Virtual machine backups taking too long to complete

If a virtual machine contains several direct-attached iSCSI LUNs or pass-through LUNs, and SnapDrive

for Windows is installed on the virtual machine, the virtual machine backup can take a long time.

The Hyper-V writer takes a hardware snapshot of all the LUNs in the virtual machine using the SnapDrive

for Windows VSS hardware provider. There is a Microsoft hot fix that uses the default system provider

(software provider) in the virtual machine to make the snapshot. As a result, the Data ONTAP VSS

hardware provider is not used for snapshot creation inside the child OS and the backup speed increases.

For more information on the Microsoft hot

fix, see Knowledge Base article 975354 on the Microsoft

support site.

Related information

Knowledge Base article 975354 - http://support.microsoft.com/

This is from the SMHV release notes

Regards,

Richard

[christoph]

Hi Richard,

We've the same issue with a customer running SMHV on a Win2008R2 SP1 machine. SP1 does contain the KB Artikel 975354 Hotfix and the VSSAPI.DLL has a higher version number than in the article.

You know anything about the issue with the hotfix installed? We've a SQL Server running with iSCSI Guest connections. Is the registry hack with setting VSSRequestor necessary to get such a situation running?

BG Christoph