Backup ok, Verify "failed" NAS

[Gerhard Wegl]

Dear all!
We cannot restore from a backup which is stored as followes:
PVE 8.0.4 with PBS 3.0.4 on PVE (upgraded from pve 6.4.x)
NAS is per NFS connected to PVE(with gui) at: /mnt/pve/NAS2021 (NFS4)
Backup is always ok. But verify "failed" therefore the following message appears when we try to open a File Restore:
"mounting 'drive-scsi1.img.fidx/part/2' failed: all mounts failed or no supported file system (400)"
When i try to restore a complete VM:
"restore failed: reading file "/mnt/pve/NAS2021/.chunks/023a/023a71a3f3c03ffafae3606668a7860849e43b334b1afbac95a0eeef878d020c" failed: No such file or directory (os error 2)"
So, this file is really not existing. Ist this part of the index? or ist this part of the backup?
If it part of the index- can it be recreated?
If not, i cannot understand: why does the backup need about 2h and the backup-Status is "successful"
And the space is reduced afterwards.
Many thanks for help!
Regards,
Gerhard

 

[Chris]

"restore failed: reading file "/mnt/pve/NAS2021/.chunks/023a/023a71a3f3c03ffafae3606668a7860849e43b334b1afbac95a0eeef878d020c" failed: No such file or directory (os error 2)"

Seems like at least one chunk is missing. Did you already run a verify job [0] on all backup snapshots?

Snapshots from subsequent backup runs can reuse chunks indexed in the previous backup, without checking if they are already present at the server. If your previous backup snapshot fails, all chunks will be uploaded again. In that case, if the chunk still exists on your source VM, it can "heal" your previous snapshots.

So, this file is really not existing. Ist this part of the index? or ist this part of the backup?

A chunk is part of the backup snapshot, it's digest (used to reference the chunk file) is stored in the index file.

If not, i cannot understand: why does the backup need about 2h and the backup-Status is "successful"
And the space is reduced afterwards.

What space is reduced? If you do not run garbage collection, no chunks will get removed, even if not in use anymore. Did you run a garbage collection recently? Maybe check that your storage correctly updates access timestamps, see [1]

[0] https://pbs.proxmox.com/docs/maintenance.html#verification
[1] https://pbs.proxmox.com/docs/maintenance.html#garbage-collection

 

[Gerhard Wegl]

Seems like at least one chunk is missing. Did you already run a verify job [0] on all backup snapshots?

Snapshots from subsequent backup runs can reuse chunks indexed in the previous backup, without checking if they are already present at the server. If your previous backup snapshot fails, all chunks will be uploaded again. In that case, if the chunk still exists on your source VM, it can "heal" your previous snapshots.


A chunk is part of the backup snapshot, it's digest (used to reference the chunk file) is stored in the index file.


What space is reduced? If you do not run garbage collection, no chunks will get removed, even if not in use anymore. Did you run a garbage collection recently? Maybe check that your storage correctly updates access timestamps, see [1]

Sorry, the left space was reduced, i would say...

Thanks for your quick replay!

I'll check the links now...

[0] https://pbs.proxmox.com/docs/maintenance.html#verification
[1] https://pbs.proxmox.com/docs/maintenance.html#garbage-collection