Bring up a Sample K8s cluster with Containerd

No Docker

Setup

We have one master and (optional) workers.

Make all firewall ports Open. You don’t want to get stuck here.

Main Reference https://kubernetes.io/docs/setup/production-environment/tools/kubeadm/install-kubeadm/

Step 1- Install Containerd in all 4 Nodes — No docker needed

Follow this in all your 4 nodes https://kubernetes.io/docs/setup/production-environment/container-runtimes/#containerd

We will configure systemd cgroups driver for containerd.

Note that there is a bug with the /etc/containerd/config.toml containerd-config-bug

It may be apparent in the next steps when you try to do kubeadm init or in the worker nodes kubeadm join.

The workaround is to delete this file and restart the containerd service and after that the kubelet service.

rm /etc/containerd/config.toml
systemctl restart containerd
systemctl restart kubelet

Step 2 Install kubelet,kubeadm and kubectl in all nodes

Install ‘kubelet kubeadm kubectl in all 4 nodes as is written here

https://kubernetes.io/docs/setup/production-environment/tools/kubeadm/install-kubeadm/

Step 3 — In Master — do Kubeadm init and CNI plugin

Before that

• Since we are using containerd we need to specify that
• Since we are using Flannel , we need a specific pod cidr to be specified

Create a config.yaml with the following

apiVersion: kubelet.config.k8s.io/v1beta1
kind: KubeletConfiguration
cgroupDriver: systemd
---
apiVersion: kubeadm.k8s.io/v1beta2
kind: ClusterConfiguration
networking:
  podSubnet: "10.244.0.0/16" # --pod-network-cidr

In Master Node

kubeadm init --config config.yaml

Note — If you are getting an error as specified in the issue containerd-config-bug, do the workaround specified

On success, you should get

Your Kubernetes control-plane has initialized successfully!
To start using your cluster, you need to run the following as a regular user:

Do as it is printed

Next is POD network — we will use Flannel

• Apply Container Networking in Master node

kubectl apply -f https://raw.githubusercontent.com/coreos/flannel/master/Documentation/kube-flannel.yml

Step 4 — In all workers do kubeadm join

Note — If you are getting an error as specified in the issue containerd-config-bug, do the workaround specified

In case you find problems with node not joiining and gettting the error CNI Network not available do the workaround

Step 5 — Copy kubeconfig to your local machine

Your cluster should be up. Copy the kubeconfig from the master node (cat ~/.kube/config ) to your lapop

Since we have not configured SSL set the insercure option via kubectl in your laptop for easy working.

kubectl config set-cluster kubernetes-green --insecure-skip-tls-verify=true

That’s it.