Moving VMs/LXCs to New Cluster

[Allister]

Hi there -

I have a current 3 node cluster that is made up of a mix-match set of machines. I'm in the process of building a new cluster (3 nodes) with identical hardware. I was thinking that it would be easiest if I just made the new cluster, backed up the VMs and LXCs, and then restored those backups in the new cluster.

Would this be the cleanest way to move everything from the old cluster to the new one?

Thank you for your time!

 

[UdoB]

Would this be the cleanest way to move everything from the old cluster to the new one?

For me there are two valid approaches:

1) use your existing backup destination to restore the VMs on the new independent cluster

2) add those new systems to the old cluster, migrate all VMs and remove the old nodes. You see, I can not know why you want to instantiate a new cluster from scratch...

Whatever makes more sense to you...

 

[Allister]

For me there are two valid approaches:

1) use your existing backup destination to restore the VMs on the new independent cluster

2) add those new systems to the old cluster, migrate all VMs and remove the old nodes. You see, I can not know why you want to instantiate a new cluster from scratch...

Whatever makes more sense to you...

Hey - thanks so much for the reply - I really appreciate it! Let me answer your questions - maybe it will help educate myself more:

My first cluster was made up of an Intel NUC 13 Pro, and two old (but different) MacBook Pros. Things have been working fine. I decided to get two more NUC 13 Pros with identical configs so I would have three nodes exactly the same. Not having a lot of experience with Proxmox yet, I thought that I would take the one NUC I have offline, update its firmware, and install Proxmox from scratch with the other NUCs thinking it would be a fresh install and help elevate any issues that may come up from removing and adding other nodes to an existing cluster.

One thing that I'm changing about NUC #1 is that I'm adding a small dedicated boot drive so that the VMs will have the NVMe all to themselves. This will require me to format and reinstall everything on NUC #1. So I guess I could do that, re-add it to the cluster, and remove the old node once everything is redone.

I'm fine with using the same cluster - for some reason (maybe Windows did this to me over the years), I thought that a fresh install would be the best way to go and not leave a mess behind that I didn't know how to clean up. I could make a new cluster, but sounds like that really won't gain me anything. If I do all this on the same cluster, will I have all sorts of dirty things on the drives in the cluster from the deleted nodes?

Thoughts?

 

[UdoB]

I thought that a fresh install would be the best way to go and not leave a mess behind

Sure. That idea is not wrong.

If I do all this on the same cluster, will I have all sorts of dirty things on the drives in the cluster from the deleted nodes?

No. As long as you follow the docs that should not happen. ( https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_remove_a_cluster_node )

Have fun!

 

[Allister]

Cool, thanks - I'll take a look at that. I'll post back once I get this done next week at some point.

 

[gfngfn256]

Make sure you have backups of everything: All VMs & LXCs (sometimes its easy to miss one - especially one that you never have to deal with).

Also check any docs/notes you may have made (or should have!) of any changes you made to the PVE host itself; incl. disk mounts, storage configs, shares, CRON jobs, fstab settings, UPS client/server configs etc. You'll come to miss these later - in afresh install!

Good luck & happy Proxmoxing.