Issue After Proxmox Update – VM Migration Fails for VLAN-Tagged Interfaces

Dec 16, 2025
6
0
1
Hello,


Today I performed a regular update on my Proxmox node. After the update, I am facing an issue while migrating VMs from one node to another.


The migration fails specifically for VMs that have a VLAN tag configured on the network interface. VMs without VLAN tagging migrate successfully.


Below is the error observed during migration:

2025-12-16 15:08:27 use dedicated network address for sending migration traffic (192.168.130.121)
2025-12-16 15:08:27 starting migration of VM 1008 to node 'prox-node4-dc2' (192.168.130.121)
2025-12-16 15:08:27 found generated disk 'data_storage:vm-1008-cloudinit' (in current VM config)
2025-12-16 15:08:27 found local disk 'data_storage:vm-1008-disk-0' (attached)
2025-12-16 15:08:27 copying local disk images
2025-12-16 15:08:28 full send of data_storage/vm-1008-cloudinit@__migration__ estimated size is 421K
2025-12-16 15:08:28 total estimated size is 421K
2025-12-16 15:08:28 TIME SENT SNAPSHOT data_storage/vm-1008-cloudinit@__migration__
2025-12-16 15:08:28 successfully imported 'data_storage:vm-1008-cloudinit'
2025-12-16 15:08:28 volume 'data_storage:vm-1008-cloudinit' is 'data_storage:vm-1008-cloudinit' on the target
2025-12-16 15:08:28 starting VM 1008 on remote node 'prox-node4-dc2'
2025-12-16 15:08:31 [prox-node4-dc2] no physical interface on bridge 'vmbr0'
2025-12-16 15:08:31 [prox-node4-dc2] kvm: -netdev type=tap,id=net0,ifname=tap1008i0,script=/usr/libexec/qemu-server/pve-bridge,downscript=/usr/libexec/qemu-server/pve-bridgedown,vhost=on: network script /usr/libexec/qemu-server/pve-bridge failed with status 65280
2025-12-16 15:08:31 [prox-node4-dc2] start failed: QEMU exited with code 1
2025-12-16 15:08:31 ERROR: online migrate failure - remote command failed with exit code 255
2025-12-16 15:08:31 aborting phase 2 - cleanup resources
2025-12-16 15:08:31 migrate_cancel
2025-12-16 15:08:34 ERROR: migration finished with problems (duration 00:00:08)
TASK ERROR: migration problems

Environment details:


  • Proxmox cluster with bonded interfaces
  • vmbr0 is attached to a bond interface
  • Affected VMs are using vmbr0 with VLAN tagging
  • Migration works for non–VLAN-tagged VMs

The error indicates that the target node reports “no physical interface on bridge vmbr0”, even though the bridge is correctly linked to the bond interface.


Could you please confirm if this is a known issue or bug introduced in the latest Proxmox update? Additionally, any recommended workaround or configuration change would be appreciated.
 
It's currently not yet merged, it has only been created two hours ago, but the fix should be merged and packaged quite soon, considering the nature of the problem.
 
May I know when this patch is expected to be officially released?


I have applied the patch on my test server and VM startup over the bond interface is now working. However, I am seeing the following message during VM start:

generating cloud-init ISO
Use of uninitialized value in string eq at /usr/share/perl5/PVE/IPRoute2.pm line 76.
TASK OK
 
After update (and restart, prior to restart was ok) my proxmox host lost completely connectivity with exceptions of running VM - they use same bridge. Also internal communication host - VM stopped working (https://forum.proxmox.com/threads/network-failure-after-recent-update.178093/)
Would that fix solve it? And how to apply it since I have only ipmi access to proxmox...?
 
I have updated and modified the lines as described in the patch within the file:


/usr/share/perl5/PVE/IPRoute2.pm


After applying the changes, I restarted both pvedemon and pveproxy, and the issue was resolved.


I also rebooted the node to verify if the issue would reappear or cause any other problems, and everything is working fine.


Could you please confirm if you have applied the same code changes in the mentioned file on your side?