Upgrades on Support Tools

Fixing Helm Upgrades After Kubernetes v1.25: The PDB API Version Trap

Thu, 16 Apr 2026 09:00:00 -0500

During a recent production cluster upgrade from v1.23 to v1.25, we hit a brick wall with Helm that temporarily halted our maintenance window. All our chart updates were properly tested, our backups were verified, and our rollback plan was ready—but Helm stubbornly refused to upgrade critical workloads with a cryptic error about PodDisruptionBudgets. Here’s what happened, how we solved it, and the important lessons that will save you considerable headaches in your future upgrades.

Kubernetes Double Jump Upgrades: Skipping Versions Safely

Fri, 29 Nov 2024 12:30:00 -0500

Upgrading a Kubernetes cluster can be challenging, especially with its fast release cycle. However, double jump upgrades, where you skip intermediate versions, can help reduce downtime and effort—if done correctly. This guide explains Kubernetes’ version skew policy, API deprecation rules, and the steps to safely perform double jump upgrades.

Upgrading Kubernetes Versions: Best Practices and Considerations

Fri, 22 Nov 2024 17:00:00 -0500

Upgrading a Kubernetes cluster is essential for maintaining security, gaining access to new features, and ensuring compatibility with modern workloads. However, it requires meticulous planning and execution to avoid disruptions. This guide provides best practices and considerations for upgrading Kubernetes clusters, whether self-managed or hosted.

The Importance of Sequential Kubernetes Version Upgrades

Fri, 25 Oct 2024 14:30:00 -0500

Upgrading Kubernetes clusters requires careful planning. While skipping versions (e.g., going from 1.28 to 1.30) may seem tempting, it can introduce unexpected issues. This post explains why sequential upgrades (1.28 → 1.29 → 1.30) are essential and offers tips for safe upgrade paths.

Why Sequential Upgrades Matter

Kubernetes releases new minor versions approximately every three months, and each version contains new features, bug fixes, and deprecations. Skipping versions may create:

Kubernetes Master Class: A Seamless Approach to Rancher & Kubernetes Upgrades

Mon, 14 Oct 2024 17:30:00 -0500

Kubernetes Master Class: A Seamless Approach to Rancher & Kubernetes Upgrades

Upgrading your Rancher and Kubernetes clusters requires a strategic approach to minimize downtime and ensure seamless operation. This guide covers rules, backup operations, planning steps, and upgrade procedures for Rancher and Kubernetes environments using RKE1 and RKE2.

High-Level Rules for Upgrades

Create an upgrade plan
- Using the rule listed below, create a plan for all of your upgrades and follow it.
- You might need to do multiple upgrades to get to the latest version.
- Please use Rancher Upgrade Tool to help you pick the right version(s) to upgrade to. Rancher Upgrade Tool
Don’t Rush Upgrades:
- Allow at least 24 hours between upgrades to ensure the stability of each component.
- Give yourself plenty of time for backups, testing, monitoring, and rollback. I recommend at least 4hr change window for each upgrade.
Don’t Stack Upgrades:
- Avoid upgrading Rancher, Kubernetes, and Docker/Containerd in one session to reduce risk. Perform them sequentially.
Backups Are Mandatory:
- Take ETCD snapshots and use the Rancher Backup Operator to ensure quick recovery in case of failure.
Upgrade Order:
- Follow the sequence: Rancher → Kubernetes → Docker/Containerd → Operating System.
Pause CI/CD Pipelines:
- Halt pipelines using the Rancher API to prevent conflicts during upgrades.
Test in Non-Production Environments:
- Always validate upgrades in a lab environment before deploying to production.
Review Release Notes and Support Matrix:
- Check the Rancher release notes and Kubernetes support matrix to avoid issues with version incompatibilities.
Monitor and Verify:
- Continuously monitor the health of nodes and pods after each upgrade to ensure everything is running smoothly.