migration aborted

[ggzengel]

It's not possible to move VHDs between different storages while migration.

DRBD to ZFS:
root@px2# qm migrate 204 px5 --migration_type insecure --targetstorage ZFS_local --with-local-disks --force
2022-12-17 18:12:36 ERROR: migration aborted (duration 00:00:00): storage 'ZFS_DRBD_12' is not available on node 'px5'
migration aborted

for testing, wrong storage name:
root@px2# qm migrate 204 px5 --migration_type insecure --targetstorage ZFS_DRBD_44 --with-local-disks --force
storage 'ZFS_DRBD_44' does not exist

DRBD to DRBD:
root@px2# qm migrate 204 px5 --migration_type insecure --targetstorage ZFS_DRBD_45 --with-local-disks --force
2022-12-17 18:10:42 ERROR: migration aborted (duration 00:00:00): storage 'ZFS_DRBD_12' is not available on node 'px5'
migration aborted

pveversion
pve-manager/7.3-3/c3928077 (running kernel: 5.15.74-1-pve)

root@px2# cat /etc/pve/storage.cfg
dir: local
disable
path /var/lib/vz
content vztmpl,snippets,images,rootdir,iso
prune-backups keep-all=1
shared 0

zfspool: ZFS_local
pool zpool1/proxmox/local
blocksize 16k
content images
mountpoint /srv/zfs/proxmox/local
nodes px5,px4,px2,px1
sparse 0

drbd: ZFS_DRBD_12
resourcegroup zfs_12
content images
controller px1
nodes px1,px2

drbd: ZFS_DRBD_45
resourcegroup zfs_45
content images
controller px4,px5,px6
nodes px4,px5

nfs: ZFS-High-Backup
export /store/zmt-backup/ShortTerm/px-images/high
path /mnt/pve/ZFS-High-Backup
server zfs1
content backup
options vers=4.2
prune-backups keep-last=2

nfs: ZFS-Low-Backup
export /store/zmt-backup/ShortTerm/px-images/low
path /mnt/pve/ZFS-Low-Backup
server zfs1
content backup
options vers=4.2
prune-backups keep-last=2

glusterfs: GFiles
path /mnt/pve/GFiles
volume gluster1
content iso
prune-backups keep-all=1
server px4
server2 px5

root@px2# cat /etc/pve/nodes/px2/qemu-server/204.conf
agent: 1
boot: c
bootdisk: scsi0
cores: 2
cpu: host
memory: 8192
name: ws-v-fax
net0: virtio=92:96:B1:ED:69:37,bridge=vmbr0001
numa: 0
onboot: 1
ostype: win10
scsi0: ZFS_DRBD_12:vm-204-disk-1,cache=writeback,size=50G
scsihw: virtio-scsi-pci
smbios1: uuid=cc1d17a3-7a2f-4367-ba12-8bba3c985eac
sockets: 2
tablet: 0

 

[ggzengel]

It's fully ignoring targetstorage:

root@px2# qm migrate 204 px1 --migration_type insecure --targetstorage ZFS_local --force
2022-12-17 19:59:02 starting migration of VM 204 to node 'px1' (172.19.80.101)
2022-12-17 19:59:03 migration finished successfully (duration 00:00:01)
root@px2# ssh px1 qm migrate 204 px2 --migration_type insecure --targetstorage ZFS_local --force
2022-12-17 19:59:42 starting migration of VM 204 to node 'px2' (172.19.80.102)
2022-12-17 19:59:43 migration finished successfully (duration 00:00:02)

root@px5# qm start 204 --migratedfrom px2 --migration_type insecure --targetstorage ZFS_local
storage 'ZFS_DRBD_12' is not available on node 'px5'

 

[ggzengel]

If I move disk to ZFS_local I will get:

root@px1# qm migrate 204 px5 --migration_type insecure --targetstorage ZFS_DRBD_45 --with-local-disks
2022-12-17 20:22:54 starting migration of VM 204 to node 'px5' (172.19.80.205)
2022-12-17 20:22:54 found local disk 'ZFS_local:vm-204-disk-1' (in current VM config)
2022-12-17 20:22:54 copying local disk images
2022-12-17 20:22:54 ERROR: storage migration for 'ZFS_local:vm-204-disk-1' to storage 'ZFS_DRBD_45' failed - cannot migrate from storage type 'zfspool' to 'drbd'
2022-12-17 20:22:54 aborting phase 1 - cleanup resources
2022-12-17 20:22:54 ERROR: migration aborted (duration 00:00:01): storage migration for 'ZFS_local:vm-204-disk-1' to storage 'ZFS_DRBD_45' failed - cannot migrate from storage type 'zfspool' to 'drbd'
migration aborted

 

[fiona]

Hi,
yes, this is a limitation/bug and was discussed here: https://bugzilla.proxmox.com/show_bug.cgi?id=3229 Currently, migration ignores the targetstorage option for storages that are considered shared like DRBD, even if they are restricted to certain nodes. Unfortunately, I haven't come around to it yet, and I also wanted to avoid conflicts with the ongoing remote-migration work, but that has now landed.

 

[fiona]

2022-12-17 20:22:54 ERROR: storage migration for 'ZFS_local:vm-204-disk-1' to storage 'ZFS_DRBD_45' failed - cannot migrate from storage type 'zfspool' to 'drbd'

And this error is because the storages have no common export/import format for offline migration. The ZFS plugin currently only supports exporting/importing via ZFS and not to other storage types.

 

[ggzengel]

This is a big penalty while migrating from old to new hardware,
We installed a new Linstor controller with a new clean database one the new hardware.

I want migration with dd (simple and stupid blockcopy), even the sources are shared (like --force_dd).