Moving Proxmox to a New Server

[Itay1778]

Hi, everyone.
So I've upgraded to a new server, and it's time to move the PVE install (1 of 2) to the new server
The storage that contains the VM and CT will also move to the new server

Of course, I'll back up everything both to a network drive and via PBS
But because I'm moving the storage with me, I think backing up and restoring to the same storage will simply be a waste of time and resources.
The question is, how else do I transfer it?
I'll reinstall PVE. One of the reasons is that I'm moving to a system with UEFI.
When I connect the existing storage to the new installation, will the LVM disks be recognized in the new installation? Or do I need to reconfigure it there, or is there a way to copy the config of that?
For ZFS, just do a zpool import?
And how do I transfer all the VMs and CTs, and the tags system?
I thought about copying their conf files, is there a better way?
The server I'm replacing is currently part of a cluster. How do I perform this transfer?
Before the move, do I remove it from the cluster?
Are there other things I need to consider?
I'll probably have to redo all the changes I made to the host - I just need to remember what I did

 

[alexskysilk]

I'll reinstall PVE. One of the reasons is that I'm moving to a system with UEFI.

You dont actually need to. As long as your host allows legacy boot you can just boot your existing installation and convert to to UEFI using proxmox-boot-tool. rest of the answers will apply to a fresh install

When I connect the existing storage to the new installation, will the LVM disks be recognized in the new installation? Or do I need to reconfigure it there, or is there a way to copy the config of that?

Yes, your new OS install will recognize the VGs. as long as your storage is configured the same (note /etc/pve/storage.cfg) it'll just work.

For ZFS, just do a zpool import?

Yes.

And how do I transfer all the VMs and CTs, and the tags system?

safest way- restore them from PBS. otherwise just transfer a copy of the original /etc/pve/qemu-server and /etc/pve/lxc

The server I'm replacing is currently part of a cluster. How do I perform this transfer?

This gets tricky since the above questions imply you dont have shared storage. Without shared storage you'd need to remove the assets from the cluster as well as the host. Otherwise just follow the instructions for node removal: https://pve.proxmox.com/wiki/Cluster_Manager#_remove_a_cluster_node and then add as new.

Are there other things I need to consider?

Yes. the interfaces file will be different unless its identical hardware, so you will need to handcraft it to make sure the bridges are functional the way you intended.

 

[Itay1778]

You dont actually need to.

Yes, I know, but right now, where Proxmox is installed is the only disk I'm not moving.
So that means cloning the disk, and that's an option I've done in the past. But I'm taking the opportunity to start from scratch.

This Proxmox installation has had about 4 major updates
And while it still works fine, I want to start fresh to clear all the junk that I've made over the years.

as long as your storage is configured the same (note /etc/pve/storage.cfg) it'll just work.

So I do need to copy this file?

This gets tricky since the above questions imply you dont have shared storage.

I do have shared storage, but I used it only for HA VM and CT, and ISO storage. The rest is in node storage

the interfaces file will be different unless its identical hardware, so you will need to handcraft it to make sure the bridges are functional the way you intended.

Yes, I know. But thanks for the reminder!

 

[alexskysilk]

So I do need to copy this file?

No. It will be reapplied when you join the cluster.

I WOULD recommend moving all your assets to your shared storage before proceeding. it will make the whole process much easier and less prone to error. You can move them back when you're done (or better yet, dont bother with the local storage at all.)

 

[Itay1778]

No. It will be reapplied when you join the cluster.

I WOULD recommend moving all your assets to your shared storage before proceeding. it will make the whole process much easier and less prone to error. You can move them back when you're done (or better yet, dont bother with the local storage at all.)

How will it be reapplied? This file is node-specific and not at the datacenter level.

I would move to full share storage, but currently, I don't have the network bandwidth to run that.
And moving all the VMs and CTs to share storage only for this swap is a problem because I don't have enough fast storage on my storage server. And on the slow storage, it will take much longer. I want to minimize the downtime.

 

[alexskysilk]

How will it be reapplied? This file is node-specific and not at the datacenter level.

everything in /etc/pve is cluster wide.