Restart a server which was running on a node that crashed on an other node

[valferon]

Hello,

I am looking for the easiest way to get back working a VM wich was running on a node of my Proxmox cluster.

My configuration is :

One master / one slave
DRBD configured in Primary/Primary mode

What is do for this test is :

-"halt" on the slave node

The only thing I have is a logical volume, replicated with DRBD on my master
=> LV : vm-101-disk-1

The only way I found so far is to create an empty VM on the master.
then I delete the disk of this new VM.

Finally I do : qm set --ide1 /dev/drbdvg/vm-101-disk-1

I can start my server on my other node.

Is there a better solution to have the same result?

Thanks for making such a good product!

 

[valferon]

I create an empty VM because, even if the original VM disk is replecated with DRBD, the conf file in /etc/qemu-server/ is not replicated.

BTW, if you have a solution to get this LVM Logical Volume and to move it to an other machine (I mean not proxmox) and to mount it, I'm interested!

 

[udo]

Hi,
simply rsync the configs to an backup-dir:
on host-a:
rsync /etc/qemu-server/ host-b:/backup/config-host-a

on host-b vice versa.
If one node fail - copy the config ( e.g. cp /backup/config-host-b/110.conf /etc/qemu-server ) and start the VM.

for rsync is cron your friend.

Udo

 

[valferon]

Ok, easy solution ^^ Thanks Udo!