VM backup fleecing doesn't work with storage=drbd

[fdlab76]

When I use the "fleecing" feature for backup I get the following error

INFO: starting new backup job: vzdump 100 --storage QNAP --compress zstd --fleecing '1,storage=drbd_0' --notes-template '{{guestname}}' --mode snapshot --node pve-2 --mailnotification always --all 0 --mailto INFO: Starting Backup of VM 100 (qemu) INFO: Backup started at 2024-07-12 22:05:40 INFO: status = running INFO: VM Name: Debian INFO: include disk 'scsi0' 'drbd_1:pm-098df180_100' 33555416K INFO: backup mode: snapshot INFO: ionice priority: 7 INFO: creating vzdump archive '/mnt/pve/QNAP/dump/vzdump-qemu-100-2024_07_12-22_05_40.vma.zst' ERROR: allocated name ('vm-100-fleece-0') has to be a valid UUID, legacy, state, or cloud-init name at /usr/share/perl5/PVE/Storage/Custom/LINSTORPlugin.pm line 385. INFO: aborting backup job INFO: resuming VM again ERROR: Backup of VM 100 failed - allocated name ('vm-100-fleece-0') has to be a valid UUID, legacy, state, or cloud-init name at /usr/share/perl5/PVE/Storage/Custom/LINSTORPlugin.pm line 385. INFO: Failed at 2024-07-12 22:05:40 INFO: Backup job finished with errors INFO: notified via target TASK ERROR: job errors

Configuration:

Proxmox 8.2.4

linstor-client 1.23.0-1 linstor-common 1.28.0-1
linstor-controller 1.28.0-1 linstor-proxmox 8.0.3-1
linstor-satellite 1.28.0-1
python-linstor 1.23.0-1

[rck]

did you trigger the backup job from the GUI? where can I find the "fleecing" checkbox? I assume this is a VM on DRBD. and then you click on the "backup" tab for that VM and then? what kind of storage do you select and where would the "fleecing" checkbox/option be? screenshots welcome

[rck]

okay, found it, it had to be a "backup job" in the "datacenter", not a direct backup of a VM :woozy_face:

[rck]

can you confirm that 264b1b71e50f4db6d6ff1af7ef05c7f6698e2b3b fixes the issue? getting the checkout and doing something like

#!/usr/bin/env bash
set -eu

[ $# -eq 1 ] || exit 1
HOST=$1

scp ./LINSTORPlugin.pm root@"${HOST}":/usr/share/perl5/PVE/Storage/Custom/
scp DRBDPlugin.pm.divert root@"${HOST}":/usr/share/perl5/PVE/Storage/DRBDPlugin.pm
ssh root@"${HOST}" mkdir /usr/share/perl5/LINBIT 2>/dev/null || true
scp ./LINBIT/Linstor.pm root@"${HOST}":/usr/share/perl5/LINBIT/Linstor.pm
scp ./LINBIT/PluginHelper.pm root@"${HOST}":/usr/share/perl5/LINBIT/PluginHelper.pm
ssh root@"${HOST}" "systemctl restart pvedaemon"
echo "OK: ${HOST}"

should do

[fdlab76]

Great job, works perfectly.

Thank you very much.

backup log below

==NFO: starting new backup job: vzdump 100 --all 0 --node pve-2 --mailto mail@example.net --notes-template '{{guestname}}' --mode snapshot --storage QNAP --compress zstd --fleecing '1,storage=drbd_0' --mailnotification failure INFO: Starting Backup of VM 100 (qemu) INFO: Backup started at 2024-07-15 12:47:12 INFO: status = running INFO: VM Name: Debian INFO: include disk 'scsi0' 'drbd_1:pm-098df180_100' 33555416K INFO: exclude disk 'scsi1' 'drbd_1:vm-100-disk-2' (backup=no) INFO: exclude disk 'scsi2' 'drbd_0:pm-bc57a65d_100' (backup=no) INFO: backup mode: snapshot INFO: ionice priority: 7 INFO: creating vzdump archive '/mnt/pve/QNAP/dump/vzdump-qemu-100-2024_07_15-12_47_12.vma.zst'

NOTICE Trying to create diskful resource (vm-100-fleece-0) on (pve-2). INFO: drive-scsi0: attaching fleecing image drbd_0:vm-100-fleece-0 to QEMU INFO: issuing guest-agent 'fs-freeze' command INFO: issuing guest-agent 'fs-thaw' command INFO: started backup task 'd2772034-e08a-47bf-8461-54547856d568' INFO: resuming VM again INFO: 2% (913.5 MiB of 32.0 GiB) in 3s, read: 304.5 MiB/s, write: 240.6 MiB/s INFO: 6% (2.2 GiB of 32.0 GiB) in 6s, read: 442.2 MiB/s, write: 190.0 MiB/s INFO: 23% (7.5 GiB of 32.0 GiB) in 9s, read: 1.8 GiB/s, write: 19.4 MiB/s INFO: 31% (10.1 GiB of 32.0 GiB) in 12s, read: 870.7 MiB/s, write: 135.0 MiB/s INFO: 41% (13.3 GiB of 32.0 GiB) in 15s, read: 1.1 GiB/s, write: 32.1 MiB/s INFO: 56% (18.1 GiB of 32.0 GiB) in 18s, read: 1.6 GiB/s, write: 138.8 MiB/s INFO: 65% (20.9 GiB of 32.0 GiB) in 21s, read: 959.4 MiB/s, write: 76.3 MiB/s INFO: 78% (25.2 GiB of 32.0 GiB) in 24s, read: 1.4 GiB/s, write: 60.2 MiB/s INFO: 97% (31.2 GiB of 32.0 GiB) in 27s, read: 2.0 GiB/s, write: 57.7 MiB/s INFO: 100% (32.0 GiB of 32.0 GiB) in 29s, read: 398.5 MiB/s, write: 342.0 KiB/s INFO: backup is sparse: 29.22 GiB (91%) total zero data INFO: transferred 32.00 GiB in 29 seconds (1.1 GiB/s) INFO: archive file size: 781MB INFO: adding notes to backup INFO: prune older backups with retention: keep-daily=15 INFO: pruned 0 backup(s) INFO: Finished Backup of VM 100 (00:00:55) INFO: Backup finished at 2024-07-15 12:48:07 INFO: Backup job finished successfully TASK OK

[rck]

thanks for testing. some other things accumulated, so I guess I will do a release very soon anyways. just one thing: DRBD should support that use case, sure, but I'm really not sure if it is a good idea to use a DRBD disk for that in the first place? according to my understanding this fleece-disk is like a scratchpad, and it should be local and fast. I'd guess other types would make more sense in this particular case (e.g., local lvmthin). VM data on DRBD for obvious reasons, fleece scratch space on local disk space. no?

[fdlab76]

Of course, you are completely right, but when I configured the cluster the fleecing option did not exist and I allocated all the space as DRDB and I need to resize it to create lvm-thin local storage... I need fleecing for backup on a really slow NAS and without it I have problems saving a VM with a database that writes continuously. I will definitely do this during a cluster maintenance interval. Thanks again for the support. Kind regards

[rck]

fixed in 8.0.4 https://github.com/LINBIT/linstor-proxmox/commit/b55df5145e3ea5866bb177bdca2a75d60d69856d