I have a standalone box that needs to be upgraded from 1.9 to 2.x. I kind of want some advice to reduce the chances of long down times... One VM in particular can't be down for long because it runs the phones.
But some others are huge (one in particular has around 400GB) and will take hours just to create a backup.
I have access to a spare PC capable of running Proxmox VE. It's far lower spec (dual core vs the 8 cores my real server has, 1/4 of the RAM, far smaller HDD) but it should run the 3 or so critical VMs.
My plan goes something like this:
Does that sound like a suitable plan? Is there a better method of dealing with the time taken for the big VMs to back up?
Will the cluster node freak out when the master goes missing? Clusters are a new thing for me
But some others are huge (one in particular has around 400GB) and will take hours just to create a backup.
I have access to a spare PC capable of running Proxmox VE. It's far lower spec (dual core vs the 8 cores my real server has, 1/4 of the RAM, far smaller HDD) but it should run the 3 or so critical VMs.
My plan goes something like this:
- Make cluster with original server as master and spare one as a node.
- Migrate important VMs over to the spare box one at a time, checking it can cope with the load
- Shut down least critical VMs and create backups using vzdump onto spare HDD
- Repeat above until all VMs are dealt with.
- Remove backup HDD (is that necessary? will the installer hose it if it's not /dev/sda?)
- Format and install on original drives in main server
- Take down the critical VMs, back them up and restore on main server.
- Restore less critical VMs from backup drive
Does that sound like a suitable plan? Is there a better method of dealing with the time taken for the big VMs to back up?
Will the cluster node freak out when the master goes missing? Clusters are a new thing for me