How to Free Up Repository Space?

When you delete recovery points, regardless of how you do it, (delete range, delete base image, rollup, hard delete, etc.) that marks the recovery point for deletion from the repository. You then have to wait for the background job "deleting index RPFS files" to run and actually remove the data. To see those jobs running go to the events tab -> click the gear icon to "show background service jobs" and then you will see them. If there are no actively running Deleting Index RPFS Files jobs then you need to restart the core service. That usually means that something in the delete scheduler hung or deletes were cancelled and now it is not cleaning up data. Please note that those jobs also may be blocked by other jobs. So it is very possible if you have an extremely busy core that the jobs are queued but always blocked by something else.

For future reference we don't ever recommend using hard delete. Hard delete will remove the recovery point that you hard delete without regard for the rest of the incremental chain. So if you hard delete a recovery point it will leave all the incrementals after that point as orphans and unusable in the repository. What we do use it for at times is to delete one volume image from a recovery point. So say you were backing up C and D on a system and you decide you want to keep C but remove all your backups of D. We would use hard delete to delete the base image of D and then use the delete orphans button to remove all the orphaned backups of D. That would leave you with your C drive backups and remove D.

Tim, are you aware of any way to force a rpfs delete to run? Would enabling deferred delete and then changing the start time of the nightly job work? If so, is enabling deferred delete necessary or what issues could this create? Thinking more about a repo that is already low on space and trying to get the rpfs to run first

As soon as the job that generates an RPFS file deletion finishes, that deletion should run. If you don't see that happening then the only way to initiate the RPFS deletions to start is to do a dismount and remount of the repository. Generally the fastest way to do that is to restart the core service (especially now in 6.2.x with the shutdown improvements that have been made). The other option is to run a simple repository check, but based on the size of the repo it might take a while. When a repository is mounted the RPFS deletions queued up and begin immediately.

As soon as the job that generates an RPFS file deletion finishes, that deletion should run. If you don't see that happening then the only way to initiate the RPFS deletions to start is to do a dismount and remount of the repository. Generally the fastest way to do that is to restart the core service (especially now in 6.2.x with the shutdown improvements that have been made). The other option is to run a simple repository check, but based on the size of the repo it might take a while. When a repository is mounted the RPFS deletions queued up and begin immediately.