reactive node with new motherboard

[Paolo_Bernardoni]

Hello,

I have a 4 nodes cluster with promox 3.1
Yesterday, one of the nodes’ motherboard failed, so I manually moved the config files of the virtual servers on the other nodes, entering the following command line:
mv /etc/pve/nodes/proxmox11/qemu-server/124.conf /etc/pve/nodes/proxmox13/qemu-server/124.conf and now everything works fine.

I know replaced the faulty motherboard, so I have to restore the node with the old hard drives, that are already configured, which means that in the folder /etc/pve/nodes/proxmox11/qemu-server I already have the files that I have to move yesterday.

What shall I do to avoid problem?
Shall I log in as single user to delete the config files I copied yesterday?
Shall I run everything normally and let the system handle it automatically?
Shall I log in as single user to move the files back to the restored node and reboot it in the network?

Any other idea/help?
Thanks

 

[danielmash]

Hi,

There is like a draft plan:

1) Just in case make backup of your production cluster eg. tar czf pvebackup.tgz /etc/pve && tar czf rootbackup.tgz /root
2) Disconnect new box from the network.
3) Single mode or you can perform normal boot at least if restored node is completely diconnected from all networks. You have to do it anyway to confirm if all hardware picked properly.
4) Remove, rename or disable vm config.
5) Make sure all hdd mounted, If you are using shared storage and depend on type you might need to learn about disaster recovery procedure of that particular type.
6) shutdown, connect your network cables and boot.
7) Check if restored system sucessfully joined the cluster (~# clustat), check shared storage status (if you use one) and migrate your vm.

Good luck.