Some VMs are not migrated correctly

[Skibbi]

Hello,
I have a small 3 nodes proxmox cluser with ZFS. It works file most of the time, but when I shutdown node with VMs with enabled HA mode for some of the machines migration fails with following error:

ERROR: tunnel replied 'ERR: resume failed - VM 102 not running' to command 'resume 102'

On the target host I'm seeing only something like this:

VM 102 qmp command failed - VM 102 not running

The VMs are just simple debian guests without any special settings. So far I cannot find any pattern, some VMs are not migrated correctly and some are migrated just fine. Any ideas how to debug this? I'm using latest proxmox 8.2.7 (6.8.12-2-pve kernel) with all updates installed.

 

[Skibbi]

I can't mark the thread as solved, but the root cause was that some of my VMs were set with CPU=host, and those were not able to migrate seamlessly between proxmox nodes. Setting virtual CPU solved the problem.

 

[LnxBil]

I can't mark the thread as solved

Have you tried editing the first post?

 

[Skibbi]

Have you tried editing the first post?

I don't have an option to edit it anymore