Three things to check on every node

A Proxmox node can look healthy while being behind in three ways: updates are waiting, an update was installed but a reboot has not happened yet, or the node's version has drifted from the others in the cluster. A fourth check looks inside the guests: virtual machines running an operating system that no longer receives security fixes.

  1. Compare kernelsuname -r against the newest installed
  2. Newer one installed: reboot dueboot into itSame kernel: up to datefor the kernel
A kernel update does nothing until the node boots into it. Some library updates also need a restart.

Step 1: See pending updates

Refresh the package lists first, the same as apt update, then list what is waiting. Replace pve1 with your node name. See the pvesh manual for more on this API.

pvesh create /nodes/pve1/apt/update
pvesh get /nodes/pve1/apt/update --output-format json
apt list --upgradable

Step 2: Is a reboot due?

An installed kernel update does nothing until the node boots into it. Compare the running kernel with the newest installed one:

pveversion
uname -r
proxmox-boot-tool kernel list
pveversion -v | grep proxmox-kernel

The first line of pveversion shows the running kernel. If a newer kernel is installed than the one running, a reboot is due. For example, running 6.8.12-4-pve with 6.8.12-9-pve installed means a reboot is due. This check covers the kernel; some library updates also need a restart.

Step 3: Compare versions across nodes

pvesh get /nodes/pve1/version
pvesh get /nodes/pve2/version

Nodes in one cluster should run the same Proxmox version. A difference is normal only in the middle of an upgrade.

Step 4: Check the guest operating systems

For a VM with the QEMU guest agent installed, ask what it runs:

qm guest cmd 100 get-osinfo

Compare the result with its support dates on endoflife.date. A guest past its end-of-support date no longer gets security fixes, however current the host is.

Rebooting a node without surprises

Getting this report automatically

Pontavia's Patch & Reboot Report ($39) is an n8n workflow that reports which nodes have Proxmox updates waiting, which run an older kernel than the one installed, version drift between nodes, and VMs whose operating system is past end of security support. It uses a read-only token, checks VM operating systems through the QEMU guest agent, and has an editable end-of-life table. It is also in the $89 Ops 3-Pack.

Frequently asked questions

How do I know if a Proxmox node needs a reboot?

Compare the running kernel from uname -r with the newest installed kernel from proxmox-boot-tool kernel list. If a newer kernel is installed than the one running, the node needs a reboot to use it.

Do I need to reboot Proxmox after every update?

No. Kernel updates need a reboot to take effect, and some core library updates need a service restart. Many other updates do not need either.

Should all nodes in a Proxmox cluster run the same version?

Yes, keep them aligned. Mixed versions are only expected for the short time during a rolling upgrade.

How do I check whether a VM's operating system is out of support?

Use qm guest cmd <vmid> get-osinfo to see what a VM runs (it needs the QEMU guest agent), then compare it with the vendor's support dates, for example on endoflife.date.

Questions about your own setup? Send a message or book a free 15-minute call.