Policy based vmdk job failing

[Ruska]

I have recently upgraded our BE 12.5 environment to 2010.  The main aim is to make use of the ability to backup virtual servers running on vSphere 4. I manually ran a backup job and restored it successfully from tape as a test.
 
I the created a Policy to create backup jobs with and tested the job created by the policy (the same server used in the successful manual run once backup job earlier). Using the policy for backing up the same server the job keeps failing with the following error.
 
-- The job failed with the following error: Unable to create a snapshot of the virtual machine. The virtual machine may be too busy to quiesce to take the snapshot.
 
 
I checked the resources in the box thinking it was low on memory which was not the case. In fact the box was sitting in idle. I followed the link on the error message but none of the possible causes listed applied. Backing up the same server manually without using the policy works fine leading me to believe it is some setting in the policy causing it to fail with the above error. In other words, what settings does a policy based backup include that the manual run does not. I created the policy using the policy wizard and kept all settings standard as far as I could. Are there any settings that could be enabled that produces the  error mentioned above? How can I weigh up the different sets of settings between a manual job and a policy job.

[CraigV]

Hi Ruska,

Some links to check out. See if they help or not:

http://seer.entsupport.symantec.com/docs/351262.htm

http://seer.entsupport.symantec.com/docs/352025.htm

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1007...

Laters!

[Ruska]

Thanks Craig

Some interesting material indeed with a few possible scenarios. I have checked various services but am not familiar with the BackupExec  Vss Provider or the BEVSSProvider.
 
What is confusing is that the box can snap in VMware. The box does snap and backup successfully when creating a run once job. It is only failing on creating a snapshot when it is trying to do so with a job created from a policy. The fact that the job completes successfully when not using a policy makes the policy suspect. I have cross referenced the options in the policy the same as the default VMware options in BE (which is assume is the options used when creating a run once job which completes successfully) as far as I can. The result remains the same, manual run successful, policy job fail.
 
Is there something hardcoded on the in the policy perhaps that requires specific config on the node side? A setting that it does not display in the properties of the policy perhaps?
 

[CraigV]

Weird...a policy would simply string those jobs together.
What happens if you create another policy and try backing it up again?

[Ruska]

Very wierd indeed

The first policy i created was for full weekly jobs only. No luck with that.

I just tried creating jobs from an existng policy that backs up other server. (File based bacjkups) It fails with the same error.

Next I tried creating a job from the blank default "agent for vmware" policy that comes with BE which created a full weekly and dif job. Ran the Full weekly and failed with the same error.

Everything works fine until a job is VM job is ran via policy which always fails with this same error.

[CraigV]

OK...I don't have enough experience ont he VM-BE 2010 front. I will hit the support flag on this for you...

[Ruska]

Thanks CraigV

Appreciate it, perhaps a bug on the policy side as all evidance points vm jobs by policy. Thanks for the efforts do far.

[Ruska]

Hello Craig

FYI

Still not sure what theproblem was, deleting everthing relating to the job (selction list, policy job) and redoing everything sorted it out. All is well.

I think the problem lied with the slection list as a Vm folder had a stroke in it yet no subfolders were selected for backup. Re-selecting and deselcting still did not make the selction stroke clear. I had to re create the slection list for it to go away. Looks like it got confuesed somehow and would not revive. I have seen something similar before when amending selcetion lists for folders that had been removed and the job still fails looking for the files(which have been removed from the selection side). When that happens the only way to fix it is to del the job and selection list and recreate. this must have been the same only on a VM scale.

[CraigV · Accepted solution]

Well, I have seen that resolving other issues. For some reason the selection list goes for a total ball.
Glad you got it sorted out, and you can now mark as a solution to close it off. Others will no doubt pick up on what you found out.