So, for some reason, deleting snapshots down to a bare minimum do nothing to restore free space in the bucket. I have Kopia running as a container backing up my Unraid system to Backblaze. I have a secondary backup container carrying out similar duties (Duplicati) which performs as expected to trim unwanted snapshots.
The first round of snapshot deletions was a few days ago. After a full 24 hours, nothing seemed to be deleted. So, I forced a safe full maintenance with
kopia maintenance run --full --safety=none
The maintenance ran without any error and several GBs in unreferenced blobs were deleted.
After a couple days, space was never freed up in the bucket. So, I deleted more snapshots this evening and followed it with a full maintenance run with safety turned off an hour later. Still, nothing deleted at the bucket.
I came across a similar thread where it was indicated the kopia.maintanace file was corrupted, so after searching in the config and cache folders, I can’t seem to locate it which is strange, considering there are no maintenance errors.
Any other things to look for?
I’ve read on Reddit the following:
Check out “Lifecycle Settings” of your B2 bucket. By default it keeps all the versions of all the files, even if “SYNC” Transfer Mode instructs to delete them. Switching to "Keep Only Last Version " should do the trick.
However, this sounds dangerous - can anyone confirm this is the right approach for Kopia?
EDIT: After further digging, it appears my docker container’s hostname has changed to a different 12 character string (same user root). Not sure why or when this occurred, so a different hostname is “owner”. Could this explain it? Still, no errors or complaints running maintenance with the new hostname. Should I set to the owner to the new hostname?