Upgrade Proxmox VE 8.4.1 → 9 — Step-by-step

Before you start — essential warnings

• Make full, tested backups of all VMs, containers and important host files (eg. /etc/pve, /etc/network/interfaces, /etc/resolv.conf, /etc/passwd).
• If you run Ceph (hyper-converged), upgrade Ceph to Squid (19.2) first — do not upgrade Proxmox before Ceph is compatible.
• Have console access (iKVM/IPMI/physical) available in case SSH drops during the upgrade. Use tmux or screen if you must use SSH.
• Test the whole upgrade on a non-production node first (or a VM) to validate your environment and third-party plugins.

Quick checklist (high level)

1. Verify backups and cluster health.
2. Run the built-in pre-upgrade checks pve8to9 --full.
3. Update current 8.4 packages (apt update && apt dist-upgrade).
4. Switch Debian base repos from bookworm → trixie.
5. Add Proxmox VE 9 repositories.
6. Run apt update and perform the upgrade to Trixie / PVE 9.
7. Reboot node and verify; repeat for each node in cluster, one at a time.

Step-by-step commands (run as root)

Run these commands on each node (preferably via console). Keep to one node at a time in clusters — migrate VMs off the node before upgrading it.

1) Pre-checks

# Verify current PVE version
pveversion

# Run the pre-upgrade checklist (recommended)
pve8to9 --full

# Make sure system is up-to-date on 8.4.x
apt update && apt dist-upgrade -y

2) Ensure Ceph compatibility (if applicable)

If you run Ceph on your nodes (especially hyper-converged), confirm ceph --version shows 19.2 (Squid). If not, follow the official Ceph upgrade instructions to move from Quincy/Reef → Squid before continuing.

3) Update Debian base repositories (bookworm → trixie)

# replace bookworm with trixie in the sources
sed -i 's/bookworm/trixie/g' /etc/apt/sources.list
sed -i 's/bookworm/trixie/g' /etc/apt/sources.list.d/pve-enterprise.list

# comment out any other third-party bookworm lines you cannot upgrade
# then update
apt update
apt policy

4) Add Proxmox VE 9 repository

Choose enterprise or no-subscription depending on your setup.

# enterprise (deb822 style)
cat > /etc/apt/sources.list.d/pve-enterprise.sources << 'EOF'
Types: deb
URIs: https://enterprise.proxmox.com/debian/pve
Suites: trixie
Components: pve-enterprise
Signed-By: /usr/share/keyrings/proxmox-archive-keyring.gpg
EOF

# no-subscription example
cat > /etc/apt/sources.list.d/proxmox.sources << 'EOF'
Types: deb
URIs: http://download.proxmox.com/debian/pve
Suites: trixie
Components: pve-no-subscription
Signed-By: /usr/share/keyrings/proxmox-archive-keyring.gpg
EOF

# update package lists
apt update

5) Perform the upgrade

# do a dry-run check with apt
apt -s full-upgrade

# then perform the actual upgrade
apt full-upgrade -y

# if you see a request about removing 'proxmox-ve', follow the official guidance (may require removing linux-image-amd64 first)
# example: apt remove linux-image-amd64

6) Reboot and verify

reboot

# after reboot, check versions
pveversion
lsb_release -a
# confirm kernel is at least 6.14 and PVE reports 9.x

Cluster upgrade notes

• Upgrade nodes one at a time. Migrate guests off the node, upgrade it, verify cluster health after reboot, then continue with the next node.
• HA groups are deprecated in PVE 9 and will be migrated to HA rules automatically once all nodes are upgraded.
• Ensure corosync/quorum looks healthy before and after each node upgrade.

Troubleshooting & known issues

• If the upgrade attempts to remove proxmox-ve, check for conflicts with packages like linux-image-amd64 and follow the official instructions to resolve them.
• Due to systemd/config changes, audit messages may appear during upgrade — you can stop the audit socket temporarily with systemctl disable --now systemd-journald-audit.socket.
• Third-party storage plugins must be compatible with PVE 9 — wait for vendor updates or test carefully.