Upgrading to Proxmox 8

In order to upgrade to Proxmox Virtual Environment 8 (from version 7) is not enough to run the usual apt update + apt dist-upgrade, major versions implies to also update APT sources.

Table of Contents

• Requirements
• Updating APT sources
• Upgrading the system
• Final steps
• References

Requirements

First of all a regular APT update + dist-upgrade must be run to be sure all requirements are OK.

apt update
apt dist-upgrade

Then we need to run migration wizard to check we accomplish all requirements:

pve7to8 --full

Typically we need, at least, to shutdown all running VMs and configure noout flag on Ceph (if we have Ceph), these are only warnings but it makes sense anyway to follow these recommendations.

Shutdown properly your running VMs on the node to update.

To change Ceph OSD settings just go to any PVE node, Ceph/OSD menu, then Manage global flags button, finally enable noout flag and press Apply button:

I had also to update time service in one of the older nodes to chrony service, maybe you don’t need it:

apt install chrony

Finally if you run again the checker you will see something like this:

= SUMMARY =

TOTAL:    44
PASSED:   40
SKIPPED:  4
WARNINGS: 0
FAILURES: 0

Updating APT sources

Just run a bunch of sed commands to do it to replace old Debian bullseye version by newest bookworm (including migrating Ceph version if needed, for instance from pacific to quincy):

sed -i 's/bullseye/bookworm/g' /etc/apt/sources.list
sed -i 's/bullseye/bookworm/g' /etc/apt/sources.list.d/*.list
sed -i 's/pacific/quincy/' /etc/apt/sources.list.d/ceph.list

Then update sources from new repos:

apt update

Upgrading the system

Just upgrade as usual:

apt dist-upgrade

Final steps

Don’t forget to unset Ceph OSD noout flag.

Reboot the system to apply new Linux Kernel, then start VMs in the rare case auto boot option is not enabled.

References

Of course Proxmox official guide is more complete than this post: https://pve.proxmox.com/wiki/Upgrade_from_7_to_8.