k8s

kubeadm cluster upgrades

on the first control plane node

First, we want to bump kubeadm itself.

Bump the repo to the target minor version (1.32 to 1.33, for example)

sudo vim /etc/apt/sources.list.d/kubernetes.list

Pick a target kubeadm version (most recent is usually fine; if there are point releases to bump you might have to do that first - ie 1.32.4 to 1.32.12, then to 1.33.X)

sudo apt update
sudo apt-cache madison kubeadm

Install the kubeadm update

 sudo apt install kubeadm='1.33.12-1.1'

Now we can start updating cluster components. Plan the upgrade:

sudo kubeadm upgrade plan

That’ll spit out a final version to aim for. Run that update:

sudo kubeadm upgrade apply v1.33.12

Let that spin for about 5 minutes. Now we can do kubelet/kubectl.

I’m lazy, I just update everything here

sudo apt update
sudo apt dist-upgrade

I reboot the node at this point as well.

second control plane node (and third)

For all other control plane nodes, the steps look mostly the same. The exception is we can skip kubeadm plan, and we’ll use the upgrade node command instead of upgrade apply

Bump the repo to the target minor version (1.32 to 1.33, for example)

sudo vim /etc/apt/sources.list.d/kubernetes.list

Pick a target kubeadm version (most recent is usually fine; if there are point releases to bump you might have to do that first - ie 1.32.4 to 1.32.12, then to 1.33.X)

sudo apt update
sudo apt-cache madison kubeadm

Install the kubeadm update

 sudo apt install kubeadm='1.33.12-1.1'

Now we update the cluster components:

sudo kubeadm upgrade node

Let that spin for about 5 minutes. Now we can do kubelet/kubectl.

I’m lazy, I just update everything here

sudo apt update
sudo apt dist-upgrade

I reboot the node at this point as well.

workers

Workers just need kubelet upgrades. So bump the repo, run the update, and (if lazy) reboot

sudo vim /etc/apt/sources.list.d/kubernetes.list


sudo apt update
sudo apt dist-upgrade
sudo reboot now