rke2-upgrades

Guides manual and automated upgrades and rollback of self-hosted RKE2 Kubernetes clusters.

2|Updated Jul 22, 2026
One-click install
npx skills add https://github.com/dydx/rke2-skills --skill rke2-upgrades-dydx
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: rke2-upgrades
Source: https://github.com/dydx/rke2-skills/tree/main/skills/rke2-upgrades
Command: npx skills add https://github.com/dydx/rke2-skills --skill rke2-upgrades-dydx

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Upgrading RKE2 clusters involves strict ordering rules, version skew constraints, and risky rollback procedures; this Skill provides the correct commands and decision rules so upgrades and downgrades are performed safely. ## Core Features & Use Cases - Manual Upgrades: Step-by-step commands for upgrading via the install script, binary replacement, or RPM packages, with server-first ordering and service restarts. - Automated Upgrades: Ready-to-use system-upgrade-controller Plan manifests for server and agent nodes, including channel tracking, version pinning, concurrency, and maintenance windows. - Rollback Procedures: Complete downgrade workflows for embedded etcd, SQLite, and external databases using pre-upgrade snapshots. - Use Case: You need to move a cluster from v1.32 to v1.33. The Skill tells you to snapshot etcd first, upgrade servers one at a time, then agents, and gives you the exact Plan YAML or install-script commands. ## Quick Start Ask how to upgrade my RKE2 cluster to the latest stable release and what to do if I need to roll back.

Frequently Asked Questions about rke2-upgrades

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I upgrade an RKE2 cluster to a new version?▼

Upgrade server nodes first, one at a time, then agent nodes. Re-run the install script with a pinned version or channel, replace the binary, or use yum/zypper for RPM installs, then restart rke2-server or rke2-agent.

How do I automate RKE2 upgrades with system-upgrade-controller?▼

Install the controller into the system-upgrade namespace, then create two Plan resources: one selecting control-plane nodes and one selecting agents with a prepare step that waits for the server plan. Set a channel URL or pin a version.

Can I skip a minor Kubernetes version when upgrading RKE2?▼

No. Kubernetes version skew policy requires upgrading one minor version at a time, and nothing in RKE2 or system-upgrade-controller blocks unsupported jumps. Patch versions within a minor can be applied directly.

How do I roll back an RKE2 upgrade?▼

Rollback requires downgrading the RKE2 binary and restoring a datastore snapshot taken on the old version. For embedded etcd, run rke2 server --cluster-reset with the snapshot path on the first server, then rejoin other nodes.

Should I use system-upgrade-controller on Rancher-managed clusters?▼

No. For Rancher-provisioned clusters, Rancher handles upgrades via the system agent, and for imported clusters Rancher manages the SUC deployment and plans. Manage upgrades through the Rancher UI instead.