Unable to run VM migration between 2 clusters

[ginesea]

Hi everyone!

I´m trying to migrate an offline VM from one cluster to another cluster in a different network segment. Firewall rules are correcly configured to allow all traffic, and API Token has admin permissions, but for some reason the establishment of the tunnel is failing.

If I try to run the migration via Proxmox Datacenter Manager v1.1.6, I get this error right at the begining of the process :
2026-07-23 14:31:02 remote: started tunnel worker 'XXXXX'
tunnel: -> sending command "version" to remote
tunnel: <- got reply
2026-07-23 14:31:02 local WS tunnel version: 2
2026-07-23 14:31:02 remote WS tunnel version: 2
2026-07-23 14:31:02 minimum required WS tunnel version: 2
websocket tunnel started
2026-07-23 14:31:02 starting migration of VM 145 to node 'HOST2' (HOST2.domain.local)
tunnel: -> sending command "bwlimit" to remote
tunnel: <- got reply
2026-07-23 14:31:02 found local disk 'ISCSI-DATASTORE:vm-145-disk-0' (attached)
2026-07-23 14:31:02 copying local disk images
tunnel: -> sending command "disk-import" to remote
tunnel: <- got reply
tunnel: accepted new connection on '/run/pve/145.storage'
tunnel: requesting WS ticket via tunnel
tunnel: Tunnel for /run/pve/145.storage failed - client error (SendRequest)
command 'dd 'if=/dev/pxe-archive/vm-145-disk-0' 'bs=64k' 'status=progress'' failed: got signal 13
command 'set -o pipefail && pvesm export ISCSI-DATASTORE:vm-145-disk-0 raw+size - -with-snapshots 0' failed: exit code 255

For the record:
- I have reviewed connectivity and it´s correct.
- Both source and destination LVM volumes are correctly configured in storage.conf file and are capable of storing the same types of data.
- SSH connection from source host to destination one works fine, and viceversa.
- Port 8006 is reachable too. I´m running same OS version in both hosts: pve-manager/9.2.3.
- I have tried differend destination datastores with no difference ( local LVM, iSCSI-based LVM volume, NVME-over-TCP ZFS volume...), same error.

Any idea what could be happening or where to look for the root cause?

Thanks a lot in advance for your help. I don´t know what else to try.

Ginés.

 

[fabian]

could you post the VM config as well as the journal covering the attempted migration from both sides?

 

[fabian]

https://bugzilla.proxmox.com/show_bug.cgi?id=7840

seems like there is a small race window when forwarding unix sockets..

 

[ginesea]

Hi Fabian,

Thanks for the help, you are right!

I can see the same error reported on that bug in the destination node:
Jul 24 13:07:36 DEST-SERVER pveproxy[3287501]: connect to 'unix/:/run/qemu-server/151.storage' failed: Permission denied at /usr/share/perl5/PVE/APIServer/AnyEvent.pm line 579.

Looks like the issue is the same, and you have already reproduced the error. If you still think sharing the journal logs would help let me know it and I'll anonymise them and upload them.

Thanks,
Ginés.