Migrate from local btrfs to local btrfs does not work

[Serverhamster]

I'm trying to migrate a VM from a local btrfs storage to another server with local btrfs storage.

2022-03-23 21:25:40 starting migration of VM 504 to node 'r2d2' (192.168.80.6)
2022-03-23 21:25:40 found local disk 'local-btrfs-sanstor:504/vm-504-disk-0.raw' (in current VM config)
2022-03-23 21:25:40 copying local disk images
2022-03-23 21:26:31 btrfs error: ERROR: cannot flip ro->rw with received_uuid set, use force if you really want that
send/receive failed, cleaning up snapshot(s)..
2022-03-23 21:26:32 ERROR: storage migration for 'local-btrfs-sanstor:504/vm-504-disk-0.raw' to storage 'local-btrfs-sanstor' failed - command 'set -o pipefail && pvesm export local-btrfs-sanstor:504/vm-504-disk-0.raw btrfs - -with-snapshots 0 -snapshot __migration__ | /usr/bin/ssh -e none -o 'BatchMode=yes' -o 'HostKeyAlias=r2d2' root@192.168.80.6 -- pvesm import local-btrfs-sanstor:504/vm-504-disk-0.raw btrfs - -with-snapshots 0 -snapshot __migration__ -delete-snapshot __migration__ -allow-rename 1' failed: exit code 2
2022-03-23 21:26:32 aborting phase 1 - cleanup resources
2022-03-23 21:26:32 ERROR: migration aborted (duration 00:00:53): storage migration for 'local-btrfs-sanstor:504/vm-504-disk-0.raw' to storage 'local-btrfs-sanstor' failed - command 'set -o pipefail && pvesm export local-btrfs-sanstor:504/vm-504-disk-0.raw btrfs - -with-snapshots 0 -snapshot __migration__ | /usr/bin/ssh -e none -o 'BatchMode=yes' -o 'HostKeyAlias=r2d2' root@192.168.80.6 -- pvesm import local-btrfs-sanstor:504/vm-504-disk-0.raw btrfs - -with-snapshots 0 -snapshot __migration__ -delete-snapshot __migration__ -allow-rename 1' failed: exit code 2
TASK ERROR: migration aborted

It looks like the error cannot flip ro->rw with received_uuid set comes from btrfs-clone. I retried after upgrading to kernel 5.15, but the result is the same.

 

[fabian]

seems like that got broken with the recent btrfs-progs update, and we need to handle this..

 

[Serverhamster]

There is a workaround. Live migration works. Offline migration does not.

 

[fabian]

yeah, live migration happens within qemu, not on the storage level.

 

[willprox]

i must delete snapshot ,then offline migration success

2022-05-16 10:29:52 starting migration of VM 101 to node 'pve02' (192.168.93.201)
2022-05-16 10:29:52 found local disk 'local-btrfs:101/vm-101-disk-0.raw' (in current VM config)
2022-05-16 10:29:52 copying local disk images
2022-05-16 10:37:28 15728640+0 records in
2022-05-16 10:37:28 15728640+0 records out
2022-05-16 10:37:28 64424509440 bytes (64 GB, 60 GiB) copied, 453.608 s, 142 MB/s
2022-05-16 10:37:28 82500+3238506 records in
2022-05-16 10:37:28 82500+3238506 records out
2022-05-16 10:37:28 successfully imported 'local-btrfs:101/vm-101-disk-0.raw'
2022-05-16 10:37:28 64424509440 bytes (64 GB, 60 GiB) copied, 453.64 s, 142 MB/s
2022-05-16 10:37:28 volume 'local-btrfs:101/vm-101-disk-0.raw' is 'local-btrfs:101/vm-101-disk-0.raw' on the target
2022-05-16 10:37:31 migration finished successfully (duration 00:07:39)
TASK OK

 

[tgm]

Fabian,
It has been almost a year... Any progress with correcting this issue? I read that live migration is a work-around but this is problematic when there are issues with starting a VM that has other problems that I'm trying to isolate.

 

[fabian]

I don't think so.. there's https://bugzilla.proxmox.com/show_bug.cgi?id=3873 for a related issue, I haven't found a tracker entry for the one in this thread though - could you maybe file one and I'll see that somebody takes it on!

 

[slaecker]

Hi - I stumbled upon this issue as well when trying to migrate a VM template from one to the other PVE cluster nodes.

Do we have an issue for that?