How to Upgrade Proxmox VE to the Latest Version (9.2 and 8.4 Point Releases)

To upgrade Proxmox VE to the latest point release, check that one working Proxmox repository is enabled, then run apt update followed by apt dist-upgrade on each node (or click Refresh and then Upgrade under the node's Updates panel), and reboot if a new kernel was installed. That takes a 9.x node to 9.2, the current release, and an 8.x node to the last 8 release, 8.4. Moving from 8 to 9 is a major upgrade with extra steps, covered in a separate guide.

This page used to be my Proxmox VE 8.3 upgrade guide. The process for point releases has not changed, so I have updated it for the current versions. It is the same routine I follow on my own homelab nodes.

Quick steps

pveversion                 # check where you are
apt update                 # refresh package lists, must finish without errors
apt dist-upgrade           # install all updates, never plain "apt upgrade"
pveversion                 # confirm the new version
reboot                     # only if a new kernel was installed

Point release or major upgrade?

Point releases (8.3 to 8.4, 9.1 to 9.2) come through the same repository, so no repository changes are needed. A major release changes the Debian base (Bookworm to Trixie), which is why it needs its own checker and steps.

Proxmox VE 8 is end of life. The Proxmox support table lists August 2026 as the end of support for 8.x. Updating to 8.4 is still worthwhile as the first step, but plan the move to 9.

Recent Proxmox VE releases

Source: the official Proxmox VE roadmap and release notes. Always read the Known Issues & Breaking Changes section for your target version before you upgrade.

Step 1: check your current version

pveversion
pveversion -v | head -5

pve-manager/9.1.x/... (running kernel: 6.17.x-x-pve)

In the GUI the version is shown at the top of the page and on the node's Summary.

Step 2: back up first

Point releases are low-risk, but kernel and storage updates can still surprise you. Before updating:

• Make sure recent backups of your VMs and containers exist (Datacenter > Backup, or Proxmox Backup Server).
• Keep a copy of the host configuration, especially /etc/pve and /etc/network/interfaces. My guide to backing up the Proxmox host configuration automates this.
• Have console access (IPMI, iKVM or a monitor) in case a kernel update affects the network.

Step 3: check the repositories

Updates only work if exactly one Proxmox repository is enabled and reachable. In the GUI open your node, Updates > Repositories: there should be no errors and either pve-enterprise (with a subscription) or pve-no-subscription enabled.

On the CLI:

ls /etc/apt/sources.list.d/
apt update

apt update must finish without 401 Unauthorized or is not signed errors. If you see those, the enterprise repository is enabled without a subscription. Fix that first with my guide to disabling the Proxmox enterprise repository and enabling no-subscription, which covers the .sources files used by Proxmox VE 9 and the .list files of Proxmox VE 8.

The suite must match your major version: trixie on Proxmox VE 9, bookworm on Proxmox VE 8. Do not change it to update within a major version.

Step 4: run the update

From the web interface

1. Select the node, open Updates.
2. Click Refresh. This runs apt update and lists the available packages.
3. Click Upgrade. A console opens and runs apt dist-upgrade. Review the list, press y and wait for it to finish.

From the shell

apt update
apt dist-upgrade

Use dist-upgrade or full-upgrade. Plain apt upgrade holds back packages that need new dependencies, so you can end up with a new pve-manager but an old kernel, or the other way round.

If apt asks about a modified configuration file, the safe default is to keep your current version (N) unless you know you never changed it. You can compare later with the .dpkg-dist file it leaves next to the original.

Step 5: reboot if needed and verify

A reboot is needed when a new kernel was installed. Compare the running kernel with the newest installed one:

uname -r
proxmox-boot-tool kernel list

If the list contains a newer kernel than uname -r, reboot:

reboot

Then check:

pveversion
systemctl --failed

pve-manager/9.2.x/... (running kernel: 7.0.x-x-pve)
  UNIT LOAD ACTIVE SUB DESCRIPTION
0 loaded units listed.

Finally clear your browser cache or press Ctrl+Shift+R so the web interface loads the new version, and check that your VMs, containers and storages are online.

Running VMs keep using the QEMU version they were started with. To pick up QEMU fixes, shut down and start them again (a reboot from inside the guest is not enough) or live-migrate them to an updated node.

Updating a cluster

Update one node at a time:

1. Migrate guests off the node (or shut them down), and check pvecm status shows the cluster as quorate.
2. Run apt update && apt dist-upgrade and reboot if needed.
3. Wait until the node is back online and green in the GUI, then move to the next node.

Mixed point releases (for example 9.1 and 9.2) in one cluster are fine while you work through the nodes. Proxmox VE 9.2 adds disarm-ha and arm-ha commands to pause HA fencing during planned cluster-wide maintenance, but the release notes warn that disarming HA while resources are still migrating can stall the upgrade of pve-ha-manager from versions before 5.2.4; re-arming HA lets it continue.

Known issues worth checking for 9.2

From the official release notes:

• Windows guests on Intel hosts with the host CPU type and VBS enabled could freeze briefly. The fix is in qemu-server 9.2.0 with machine version 11.0+pve2.
• Installer ISO kernel and some RAID controllers or NVMe drives in JBOD mode (megaraid_sas): fixed in later 7.0 kernels. This affects new installs from the ISO; updating an existing node pulls the fixed kernel.
• Ceph: Tentacle 20.2 is available as a stable option in 9.2, while Squid 19.2 remains supported. Updating Proxmox VE does not switch Ceph releases; moving from Squid to Tentacle is its own documented upgrade.

Troubleshooting

401 Unauthorized from enterprise.proxmox.com. See Step 3: disable enterprise or add your subscription key.

The following packages have been kept back. You used apt upgrade. Run apt dist-upgrade.

W: (pve-apt-hook) You are attempting to remove the meta-package 'proxmox-ve'! Stop and answer no. A Proxmox repository is missing or the suites are mixed. Check every file in /etc/apt/sources.list.d/ uses the same suite.

No network after rebooting into a new kernel. A kernel update can rename network interfaces. From the console, compare ip link with /etc/network/interfaces and fix the bridge port, then ifreload -a. Proxmox VE 9 includes pve-network-interface-pinning to pin names so it does not happen again. If ifreload is missing, see fixing "you need ifupdown2 to reload network configuration".

A VM with GPU passthrough stopped working after a kernel update. Boot the previous kernel from the boot menu, pin it with proxmox-boot-tool kernel pin <version>, and see my GPU passthrough guide.

dpkg was interrupted. An update was cut off. Run dpkg --configure -a, then apt dist-upgrade again.

Learn more

If you want a structured way to learn Proxmox, from installation to clustering and backups, enrol in my full Proxmox course.