One of the fundamental things I’m trying to achieve here is primarily a complete rebuild of the production Proxmox servers. Secondly, a rebuild of all VM’s on these servers. I’ve realised that perhaps I need to shift my focus directly on the Proxmox servers rather than the VM’s within these servers.
So let’s focus on the Proxmox servers for now.
I’ve currently got a NFS server I use for infrastructure so I can easily share resources (like backups and disk images) across infrastructure. This NFS server is also accessible by the standby server if mounted. So the plan is to utilise this to shift all the VM’s across to the standby server, rather than reinstall everything by hand.
- Mount the NFS Server on the standby Proxmox server
- Live backup a VM to the NFS server
- Deploy the backup to the standby server
- Shutdown the old VM
- Start the new VM
Should be a very minimal outage, particularly if we’re mindful not to make changes to a VM once it’s been backed up and before it’s migrated to the standby server.
This way, we can have an outage only for the time it takes for the VM to be deployed to the standby server.
Once a VM is deployed, it retains all it’s configuration. MAC address, DHCP will assign config as usual. No problems.
I’ve already tested this with the LDAP server. It’s like magic.
DHCPD and PfSense should be cool too, but I’ll have to manage that via a physical console. So particularly for these servers I must have the backup deployed to the standby server ready to go before shutting down the old VM.
As for when this will all occur. Well I expect this to take a couple hours to complete, and during Winter I go to the snow of weekends, so might just have to wait until the season closes!