Upgrading Nodes Manually

[ShunsukeMiyake]

I would like to know more about "Upgrading Nodes Manually" as described in the following document: https://rke.docs.rancher.com/upgrades#upgrading-nodes-manually

We are considering upgrading the Kubernetes version of our downstream cluster.
Could you advise on how we could initiate the upgrade to manually upgrade each node one at a time?

Our company uses Rancher to manage our clusters.
Since the target cluster for the upgrade was set up using Rancher's UI, we have access to the configuration (cluster.yaml) via the Rancher UI.
If we change the Kubernetes version in this configuration, the upgrade for each node would likely start automatically.

However, since we would like to upgrade each node manually, we are thinking of setting the following options as shown below.
Would this allow us to perform each node's upgrade manually?

• Drain nodes -> false
• Maximum Worker Nodes Unavailable -> 0

For example, within the cluster, we have Redis, with the master and slave located on separate nodes.
When the master begins upgrading, I believe it will take a certain amount of time for the slave to detect that it can no longer communicate with the master.
During this period, data cannot be recorded to the master, leading to a potential data loss for that time span.
To minimize data loss, we would like to control the order and timing of the upgrade—for instance, by starting the upgrade from the slave.

Thank you for reading.

[github-actions]

This repository uses an automated workflow to automatically label issues which have not had any activity (commit/comment/label) for 60 days. This helps us manage the community issues better. If the issue is still relevant, please add a comment to the issue so the workflow can remove the label and we know it is still valid. If it is no longer relevant (or possibly fixed in the latest release), the workflow will automatically close the issue in 14 days. Thank you for your contributions.

[IronBy]

We'd also like to know how this manual upgrade is supposed to work and how we can trigger it.

In our case we deploy and upgrade K8s clusters with Terraform rke provider. But to gain more control over upgrades we'd like to have the ability to upgrade certain nodes only using rke cli tool. For example, we might upgrade api and etcd nodes on day one and later upgrade worker nodes independently on another day.

Unfortunately this is not apparent how to perform this kind of upgrade with rke cli tool. Can someone please point us into the proper documentation or provide more detailed guidance.

Thank you in advance.

[SM-Aristo]

I'm also interested in this information, please. Thank you!

[rbiryla]

Please share more details on the topic. Thank you in advance!

[iotchenkoarist]

This is a very interesting topic! We are waiting for an answer!