RKE2 OS Upgrade Process
This document (000021301) is provided subject to the disclaimer at the end of this document.
Environment
Situation
Resolution
After upgrading each nodes, worker or controlplane/etcd, you should check that the node and cluster are healthy before upgrading the next node.
The steps to upgrade or patch the operating system are:
- During the OS upgrade, don't stop the RKE2-Server or RKE2-agent services.
- Drain worker nodes before upgrading them.
- Wait for the pods to be drained and scheduled to other nodes before upgrading the drained worker.
- Perform the upgrade in a rolling fashion:
- Start upgrading a single controlplane/etcd node at a time
- Upgrade one or more workers at a time (depending on the capacity within the cluster to reschedule workloads to other nodes)
- Reboot the node after the upgrade is completed.
- Uncordon the node
Draining nodes during the operating system upgrade:
Your controlplane/etcd nodes are not running with the worker role, so do not run any user workloads.
Whilst the cluster-agent scheduled on these nodes runs as Deployment of two replicas on separate nodes. As a result, you do not need to drain controlplane/etcd nodes.
Special care if Longhorn is installed:
If Longhorn is installed on the cluster, please take a look at any requirement on the longhorn documentation related to the version you are using.
Especially the part concerning the drain.
https://longhorn.io/docs/1.4.4/volumes-and-nodes/maintenance/#upgrading-kubernetes
Disclaimer
This Support Knowledgebase provides a valuable tool for SUSE customers and parties interested in our products and solutions to acquire information, ideas and learn from one another. Materials are provided for informational, personal or non-commercial use within your organization and are presented "AS IS" WITHOUT WARRANTY OF ANY KIND.
- Document ID:000021301
- Creation Date: 20-Dec-2023
- Modified Date:10-Jan-2024
-
- SUSE Rancher
For questions or concerns with the SUSE Knowledgebase please contact: tidfeedback[at]suse.com