I am currently in my environment and we are backing up OVM virtual machines.
While backing up the VMs, commvault created a snapshot and mounted it on the production OVM Manager. It truncates it after the backup is completed.
Is it possible for the snapshots not to mount on the OVM manager at all?
It actually affected production last time night. Please, help out
Best answer by MubaraqView original
I moved the licensing question over here:
You are very much correct. We have found out about all you’ve said overtime.
Below is what we did and the reports have been 100%, so far.
The OVM Manager is a physical server.
At some point during backup, there was a communication break between the proxy and the commserve.
The snapshot could not be deleted from the hosts. Which caused duplication of VMs, and occupied disk space on the host as a result.
Was there a Snapshot/Backup taken of the Proxy machine? Or any observed hang/crash on the VSA Proxy at all?
Yes, one of the administrators ran all the jobs at once defying the schedules, the proxy server crashed at some point leading to break in communication with the media agent and commserve.
This caused the already created snapshots not to be truncated which led to high resource usage on the hosts and caused downtime.
A new proxy has been provisioned, the backups are to run in batches. We recommended that the resources on the proxy should also be increased.
Apologies for delay in replying here!
A crash would certainly cause Snaps to be left behind and potentially disks still attached to the VSA proxy here.
For the Proxies, ensure that Automount (Windows) or lvm2-lvmetad (or other Linux Equivalent) is disabled. This is to ensure that the OS is not controlling the attached Guest Disks.
If you have any further occurrences of a crash, I would suggest opening up a support case also.
Thanks for the update
With OVM backup the Disk from the Snapshot would be attached to the VSA Proxy VM for backup, then cleaned up after.
Is the OVM a VM itself here? - If so, is this being backed up?
Could you explain the “impact” that the backup had on production please?
It may be worth raising a case to validate what occurred here.
Let us know the case ID once opened.