Backup of CT Failed Not enough free space available for volume 0

[t0mat000]

Aktuell können wir keine Backups mehr von einem CT machen. Die Disk ist mit ca. 800GB recht groß. Aber das Storage und der Backupserver haben mehr als genug freien Platz. Testweise wurde auch ein NFS Storage angebunden, aber dorthin kommt der selbe Fehler. Es scheint also irgendwie am Linstor DRBD zu liegen.

Das DRBD liegt auf einem ZFS Raid 1.

ERROR: Backup of VM 520440 failed - API Return-Code: 500. Message: Could not restore snapshot snap_vm-520440-disk-1_vzdump of vm-520440-disk-1 to new snap_vm-520440-disk-1_vzdump, because:
ERROR: [{"ret_code":-4611686018407202610,"message":"Not enough free space available for volume 0 of resource 'snap_vm-520440-disk-1_vzdump'.","details":"Resource: snap_vm-520440-disk-1_vzdump","error_report_ids":["65C0D13F-00000-000005"],"obj_refs":{"RscDfn":"snap_vm-520440-disk-1_vzdump"},"created_at":"2024-02-05T14:20:23.25359+01:00"}]
ERROR: at /usr/share/perl5/PVE/Storage/Custom/LINSTORPlugin.pm line 436.

Proxmox Plugin:
drbd: NVME_DRBD_R1_1
resourcegroup NVME_DRBD_R1_1
content rootdir,images
controller 188.94.24.233
nodes server01,server02

Proxmox Version: 7.4-17

[rck]

I don't think the issue is on the target/backup side, but on the "source". For backups to work, I think PVE needs an actual resource/block device it can create the backup from. Also IIRC, the plugin creates a LINSTOR snapshot, which creates snapshots on the actual devices, then the plugin is asked to create a new resource from that snapshot (this is where you run out of space), then PVE does its magic and calls into the plugin to delete the temporary resource + the LINSTOR snapshot.