problem with datastore based on zfs

[luca_merkle]

I have a PBS whose datastore in based on zfs, 4 disks, raidz1.

issues:
1. doesn't free up space after deleting old backups according to retention policy
2. the storage keeps filling up
3. can't do a garbage collection, as when I manually start it, it takes literally hours giving no signs of life. I try to stop it, and it results in an endless aborting attempt:
2023-07-19T11:03:47+00:00: starting garbage collection on store PROX4
2023-07-19T11:03:47+00:00: Start GC phase1 (mark used chunks)
2023-07-19T11:24:19+00:00: received abort request ...
2023-07-19T11:24:19+00:00: TASK ERROR: abort requested - aborting task

Where I am wrong?
Thanks

 

[dcsapak]

3. can't do a garbage collection, as when I manually start it, it takes literally hours giving no signs of life. I try to stop it, and it results in an endless aborting attempt:

check with a tool like atop during a garbage collect, it will probably show you that your disks are under full load
if you have many backups, garbage collect can take very long especially on spinning disks

 

[luca_merkle]

thank you