Upgrade PVE 6.4 to 8.2

[Reartu24]

Hi all
I started a new job and found myself in a bit of a strange situation....
the production cluster (6 nodes, with 25TB and 32VM ceph storage), at version 6.4.

At first I thought about upgrading from 6 to 7, and then making the leap from 7 to 8.

today in the warehouse I found 3 disused nodes, and I thought, why not create a new cluster with version 8.2 and gradually move the VM?

and here the problem, being software that cannot be turned off (I had thought of the classic backup and restore on the new host), there is a way to make the two clusters "talk" and pass the turned on VMs from one side to the other other?

Thank you so much everyone and have a good afternoon

 

[fiona]

Hi,

Hi all
I started a new job and found myself in a bit of a strange situation....
the production cluster (6 nodes, with 25TB and 32VM ceph storage), at version 6.4.

At first I thought about upgrading from 6 to 7, and then making the leap from 7 to 8.

today in the warehouse I found 3 disused nodes, and I thought, why not create a new cluster with version 8.2 and gradually move the VM?

and here the problem, being software that cannot be turned off (I had thought of the classic backup and restore on the new host), there is a way to make the two clusters "talk" and pass the turned on VMs from one side to the other other?

Thank you so much everyone and have a good afternoon

there is a remote-migration feature on the CLI for cross-cluster migration, but it only got introduced in Proxmox VE 7.3, so the nodes with 6.x are too old to be able to use it. If the VMs really can't be turned off, just upgrading the cluster normally is probably the best option. Having restore-tested backups is of course highly recommended.

The official upgrade documentation can be found here:
https://pve.proxmox.com/wiki/Upgrade_from_6.x_to_7.0
https://pve.proxmox.com/wiki/Upgrade_from_7_to_8
and please note the steps regarding Ceph.

 

[Reartu24]

Hi Fiona,
Thanks for your reply, I carried out the upgrade test on a test cluster and they went very well, so I'll get organized shortly and following the procedure I will also try to update the production cluster, hoping that everything goes smoothly.

Have a good day

Stefano