VSSFreezer failed to free guest, wait timeout

[tfleener]

I have been experiencing the follow warning on a few web/file VM servers sometimes, and other times the VMs backup without warnings.

The error reported is:
"Creating snapshot VSSFreezer; Failed to freeze guest, wait timeout" .

Log file shows:
[09.11.2010 00:16:03] <01> Info VSSFreezer: Frozen in 11sec
[09.11.2010 00:16:03] <01> Info Not using VMTools Quiesce.
[09.11.2010 00:16:03] <01> Info [VimApi] Create snapshot, ref "vm-9465", name "VEEAM BACKUP TEMPORARY SNAPSHOT", description "Please do not delete this snapshot. It is being used by Veeam Backup.", memory "False", quiesce "False"
[09.11.2010 00:17:07] <01> Info VSSFreezer: Unfreezing
[09.11.2010 00:17:08] <01> Error VSSFreezer: Failed to freeze guest, wait timeout at Veeam.Backup.AgentProvider.CVssControl.Unfreeze()
[09.11.2010 00:17:08] <01> Error at Veeam.Backup.Core.CVssAutoFreezer.Dispose()
[09.11.2010 00:17:08] <01> Info Task operation result: VSSFreezer: Failed to freeze guest, wait timeout
[09.11.2010 00:17:08] <01> Info Task operation note: VSSFreezer: Failed to freeze guest, wait timeout
[09.11.2010 00:17:08] <01> Info VssGAConn: Uninstalling agent


I have increased the VSSFreezer timeout value in the registry, however the error still occurs.

Although it is a warning, it is concerning, so I have a few questions I need clarification on:

1) Should this VSSFreezer timout error should not be considered a serious error ? That is, the Veeam VSS still perform a quiescence of the file system ? Are my backups/replicas consider good ?

2) Given I am using the change block feature; should my full back/or replica be considered to be in a consistent state ?

3) If they are not consitent, will the following synthetic backups/replicas be considered to be inconsistent since a prior VSSFreezer error occured ?

4) Is it best to follow VSSFreezers with a Fullbackup to get consistency back ?

5) If this was to occur on a VM transactional type servers (Exchange/AD or SQl), would it be considered more serious ?

I am concerned, on one hand it isa warning, on the other hand it seems like the OS/File System is missing the quiesecing of the files.

thanks for any insight,

Tom

[Alexey D.]

Hello Tom,

The issue is that this VMs reaches the timeout for snapshot creation. Microsoft has hard-coded setting which prevents to keep OS in frozen state for more than 60 seconds, and Veeam cannot change this. Take a look at these strings in the log:

[09.11.2010 00:16:03] <01> Info [VimApi] Create snapshot...
[09.11.2010 00:17:07] <01> Info VSSFreezer: Unfreezing

Please troubleshoot your VM - it looks like some services are doing bad and impede VSSFreezer to finish its job in time.
You may also want to check your vCenter load and try adding ESX host (where that VM resides) to backup console directly.

Concerning your other questions, if the freezer fails it may lead to inconsistent backups/replicas. The more high-transaction your server is, the more probability is to get its improper backup state, that's why it's highly recommended to process AD/SQL/Exchange/... with application-aware processing.

Also, consistency has no connection with job modes (incremental/full run), please read the sticky FAQ about B&R for more information on it. Consistency is handled by application-aware processing logic, and, for example, if you do incremental runs and during one run VSSFreezer failed - you may get inconsistent restore point.

[tfleener]

As follow up to the answer, it ended up the VCenter server had a debug option set from a previous tech support call with VMware. Since I disabled the VCenter debug logging option I have not seen any VSSFreezer failures. Thanks for pointing me in the right direction.