PVE Node "death"

[Volker Lieder]

Hi,
in our cluster, pve node05 is marked with a "?" in Proxmox Gui.
The VMS on that node are still running and available, but i even cant access per ssh or gui to that server.
Is there a way to migrate running vms to another node while the server is not available via ssh or mgmt gui?
icmp to mgmt network and storage network is still responding.

Infos:
Version 7.2-7
7 nodes
CEPH running

Best regards,
Volker

 

[UdoB]

Is there a way to migrate running vms to another node while the server is not available via ssh or mgmt gui?

Captain Obvious says: use the local/physical console to investigate...

 

[Volker Lieder]

the pyhsical console is also "dead" no further output, no messages in ipmi log

 

[Volker Lieder]

On the screen

 

[lDemoNl]

@Volker Lieder I migrated VM's from the inaccessible node ,but you have to know that process stop your VM's
You need to login in working node and move config file of VM
mv /etc/pve/nodes/node05/qemu-server/VM_ID.conf /etc/pve/nodes/node04/qemu-server/
And your VM will be migrated

 

[Volker Lieder]

Yes, to move configs i know about. The hope was that there is a way to do a livemigration in such a state.
Other plan is to shutdown instances and restart them on another node with config move.

 

[Volker Lieder]

We could shutdown the VMs, migrate the config and restart them. Powercycle of node also works, we will investigate the issue and observe the behaviour the next days.
Best regards,
Volker