Zero-touch node patching for Proxmox clusters by gyptazy.
Automate the most repetitive operational task in Proxmox: keeping cluster nodes updated. ProxPatch drains, migrates, patches, and reboots nodes in a controlled rolling fashion — no downtime, no manual intervention.
A predictable, auditable sequence. Every step is transparent and logged — no black boxes.
From homelab setups to production environments — ProxPatch is designed to be trusted.
Patches one node at a time while keeping the rest of the cluster fully operational. Guests stay online throughout the entire process.
Verifies cluster health before touching each node. Never proceeds if the cluster is in a degraded state or quorum is at risk.
No orchestration frameworks, no external databases, no API tokens. Uses only native Proxmox tools: pvesh, qm, and SSH.
Clear execution logs with timestamps at every step. Every decision is visible and auditable — you always know what ProxPatch is doing.
Intentionally minimal and transparent. Works equally well on a 2-node homelab and a 20-node production cluster.
Only reboots when genuinely required. Skips unnecessary restarts if the applied updates don't require a kernel change.
ProxPatch is not a full lifecycle manager or an HA replacement. It focuses on exactly one task and executes it with precision.
ProxPatch started as a planned feature of ProxLB (another tool by gyptazy) — a DRS-like load balancer for Proxmox clusters. However, missing API endpoints for rolling node patching and reboot orchestration made it necessary to build this as a standalone tool. Integrating workarounds into ProxLB would have introduced long-term maintenance risks. So ProxPatch was born as its own focused project.
Add the official repository via gyptazy open-source solutions and install ProxPatch with two commands. No build tools, no runtimes.
proxpatch package on only one node in the cluster. No external dependencies required — ready to run immediately.bookworm and trixie and is fully compatible with Proxmox VE 8.x and 9.x environments.
Drop the manual drain-migrate-patch-reboot routine. Let ProxPatch handle it while you focus on what matters.