[SOLVED] Ownershipfile being left in filesystem even if no backups exists

[bennetgallein]

Hey everyone,

looking for some insight how to cleanup the following situation:

I have a datastore that had backups, but all were removed. Garbage Collection ran successfully but there are still files and chunks leftover.

This is the output of the GC:

2025-06-13T11:37:53+02:00: Access time update check successful, proceeding with GC.
2025-06-13T11:37:53+02:00: Using access time cutoff 1d 5m, minimum access time is 2025-06-12T09:32:52Z
2025-06-13T11:37:53+02:00: Start GC phase1 (mark used chunks)
2025-06-13T11:37:53+02:00: Start GC phase2 (sweep unused chunks)
2025-06-13T11:37:54+02:00: processed 73% (0 chunks)
2025-06-13T11:37:55+02:00: Removed garbage: 0 B
2025-06-13T11:37:55+02:00: Removed chunks: 0
2025-06-13T11:37:55+02:00: Original data usage: 0 B
2025-06-13T11:37:55+02:00: On-Disk chunks: 1
2025-06-13T11:37:55+02:00: Deduplication factor: 0.00
2025-06-13T11:37:55+02:00: Average chunk size: 159 B
2025-06-13T11:37:55+02:00: TASK OK

Also in the filesystem I have a leftover "owner" file:

> # tree                                                                                                                                                                  ⬡ 20.19.2
.
└── vm
    └── 100
        └── owner

3 directories, 1 file

Any suggestions why these are leftover and how I can get rid of them (without manual intervention)? I dont think they should be there and are left over somehow.

 

[Chris]

Hi,

Hey everyone,

looking for some insight how to cleanup the following situation:

I have a datastore that had backups, but all were removed. Garbage Collection ran successfully but there are still files and chunks leftover.

This is the output of the GC:

2025-06-13T11:37:53+02:00: Access time update check successful, proceeding with GC.
2025-06-13T11:37:53+02:00: Using access time cutoff 1d 5m, minimum access time is 2025-06-12T09:32:52Z
2025-06-13T11:37:53+02:00: Start GC phase1 (mark used chunks)
2025-06-13T11:37:53+02:00: Start GC phase2 (sweep unused chunks)
2025-06-13T11:37:54+02:00: processed 73% (0 chunks)
2025-06-13T11:37:55+02:00: Removed garbage: 0 B
2025-06-13T11:37:55+02:00: Removed chunks: 0
2025-06-13T11:37:55+02:00: Original data usage: 0 B
2025-06-13T11:37:55+02:00: On-Disk chunks: 1
2025-06-13T11:37:55+02:00: Deduplication factor: 0.00
2025-06-13T11:37:55+02:00: Average chunk size: 159 B
2025-06-13T11:37:55+02:00: TASK OK

This one chunk remaining in the chunk store is introduced with the atime safety check, it is an all zero chunk which is inserted into the chunk store at the start of garbage collection and used to see if the atime updates required for garbage collection to work properly are honored by the filesystem. It is strongly recommended to keep this enabled. For details see:

• https://pbs.proxmox.com/docs/storage.html#tuning
• https://pbs.proxmox.com/docs/maintenance.html#garbage-collection

Also in the filesystem I have a leftover "owner" file:

> # tree                                                                                                                                                                  ⬡ 20.19.2
.
└── vm
    └── 100
        └── owner

3 directories, 1 file

Any suggestions why these are leftover and how I can get rid of them (without manual intervention)? I dont think they should be there and are left over somehow.

This has been fixed with https://bugzilla.proxmox.com/show_bug.cgi?id=3336, manual intervention should no longer be required as the group folder will be cleaned up when removing the last snapshot of a group.

You will however have to remove already per-existing empty group folders.

 

[bennetgallein]

I see, thanks!