LUN Backup and Storage vMotion

[jbarrow.viracoribt]

Let's say you have a backup job backing up an entire LUN, LUN 11. The job runs fine for a week, then someone performs a Storage vMotion on a VM that was on LUN 11 and moves it to LUN 22. Lun 22 happens to have it's own backup job deined that does the same thing, backs up the entire LUN. I would assume that the VM would be picke up by the LUN 22 backup job, but what happens to the data on LUN 11 from this VM's prior backups?

[foggy]

Yes, the VM being moved will be picked up by the corresponding job and a full backup of that VM will be performed. The VM will fall out of the original job according to the Deleted VMs retention period setting.

If you actively use svMotion in your environment, I would not advise to configure jobs by datastore as source container. It's more practical to use some other container (that would not be impacted by VM location) or select VMs individually.

[jbarrow.viracoribt]

We hardly ever storage vMotion, most VM's stay in the same place. It was just something I was curious about.

[Cokovic]

Thats exactly the way we do our backups here (per LUN). And we do alot of storage vMotion Works for years now here for about 2000 VMs.

Anyway theres one thing you should take care of if using reversed incrementals with this approach. Your fullbackup file will never shrink meaning you should perform a new active full from time to time to reclaim disk space.

[jbarrow.viracoribt]

Good info about reversed incremental. Our jobs are configured for Incremental with synthetic fulls on Sunday's.

[dellock6]

Haris, I hope for you you are not going to use Storage DRS, adding it to your backup design would transform it in a nightmare

Luca.