In my journey to establish a Highly Available Kubernetes Pi Cluster, I've successfully configured the cluster following an external etcd setup. The process involves installing Docker, setting up the Docker daemon, and installing kubeadm. Initializing Kubernetes Master Nodes, preparing certificates, and configuring Calico for networking are key steps. Troubleshooting tips include addressing refused connections and certificate expiration. To rejoin a faulty node, cordoning, draining, and generating new keys are essential. Now, I proudly own a fully operational Highly Available Kubernetes Pi Cluster.

I recently spent 70 minutes setting up a Kubernetes Cluster on Raspberry Pi using Ubuntu Server 20.04 LTS. After burning the OS image and configuring a headless setup, I updated the OS, changed the hostname, and enabled memory cgroup. The Docker installation involved setting up external storage and configuring Docker daemon. Installing kubeadm and creating the Kubernetes cluster took an additional 45 minutes. I verified the cluster status, installed networking addons (Calico), and added leaf nodes. Troubleshooting included resolving conntrack and socat issues. Overall, the Raspberry Pi Kubernetes Cluster provides full control over Docker container orchestration.