Restart a backup of VM from the console

[vburgun]

Hi,
Excuse me for my bad english.

We use NetBackup 7.5.0.3. We backup our VM (vSphere 4.1) using a dynamic query to generate backup jobs (Virtual Machine selection: Select Automatically throug query)
If a VM can not be saved and the job falls into error, we can not just do a restart the job from the activity monitor.
We can only restart the job master and then restart the backup of all VMs.
We had to create a specific policie manually declare the VM error and restart as well.

This is not very practical. Do you have another solution? Is there a future improvement of the future release of NetBackup?

Thanks

[Anonymous]

That is a good comment.

Because you are choosing VM selection via a query or via folders etc you have no hands on choice about what client to backup in a manual retry.

This should be posted as a IDEA.

What I would suggest would be a option that could be turned on/off or this is a behaviour design.

Basically you should be able to retry the master backup job AND it should ONLY retry those clients that failed. Or automatically retry within a defined number of attempts.

So in effect the retry should perform an "intelligent additional exclude" in the query selection result based on those in that instance completed successfully.

Maybe it might be a forthcoming feature. Just keep your eyes open in release notes.

Obviously the real goal with failed backups is to address why they failed with the status they did and fix that so they dont fail in future.... till the next time..

[Mark_Solutions · Accepted solution]

I have already raised this with Symantec during partner meetings and i guess that it hasn't been put right yet (i suggested it after the query was introduced in 7.1)

The official line is as follows:

Have a Re-Run policy available with the same schedules (but no open Windows) in NetBackup.

When a client fails add it to that policy and run it manually - this does actually work for Incremental backups as well as the query os made to vCenter (unlike any other sort of policy)

The reason is that after the query has been made it creates a file (xml i believe) that it uses for the client list and back run. Once the jobs have finished this is removed so it prevents you re-running a failed job

Not much help i know but do put it forward as a suggestion again to keep the pressure on