Re-balance extents in a SOBR

[akrietsch]

We have a SOBR with some extents that all match up to individual volumes on some Cisco storage servers. We are getting low on space across all of the extents, so we have purchased another Cisco storage server to add to the SOBR. My question is about our active backup jobs that go to this SOBR. I know that any newly backed up VMs would choose the new empty extents we add to the SOBR, but I believe that all existing backup chains would continue to write to the extents where they currently reside (which are getting full). This is because we are using the "Data Locality" policy. We do this because we use the per-VM backup chains and utilize the ReFS space savings, so we want an entire backup chain to be on the same extent.

What would be the best way to get some of our already backed up VMs to start writing to the new empty extents? The description of the data locality policy says that "...however, the next full backup can be created on another extent". If I add some empty extents to the SOBR, does this mean that the next time it takes full backups it will choose to write to the new extents since it sees there is free space and it generally tries to balance it all out?

If not and I am misunderstanding, is there something else we can do to migrate backup files from the current almost-full extent over to the new extents so that we do not run out of space?

[Shestakov]

Hello Alex,
Indeed new chain, starting from full backup should be written on the new extent. Unless you have a dedup storage appliance as SOBR extent, otherwise new chain will be placed to the extent where the previous backup chain resides.
Thanks

[akrietsch]

Thanks. Do I need to take active fulls, or if my backups are set up to take synthetic fulls once a week will it still place those on the new extents?

[Shestakov]

You don't need to start active full manually. Synthetically created full backup should be placed on the new extent.

[foggy]

In case there's no sufficient space on the extent for incremental backup, it will be placed to the new extent even before new full is created.

[williampu]

I originally had 2 volumes on NTFS and 2 on ReFS on a single Cisco Server. I have wiped the NTFS volumes and reformatted them as ReFS.
There is now a SOBR consisting of 4 ReFS volumes, however the 2 newer ReFS volumes are barely being used. Keep in mind that this SOBR is offsite and Backup Copy jobs are writing data to it.
Am I able to just temporarily disable the Backup Copy Jobs, shift the data around to the newer extents to manually balance the usage and do a rescan if the directory names are kept exactly the same?

[dellock6]

Hi William,
supposing you are using data locality policy, SOBR placement works for the initial plaement of the chain, and following incrementals stay in the same location. So, unless you have new VMs, that's the reason why the new extents are not used. Indeed, manually moving the chains and then rescanning the extents could do the trick.