Kubernetes Worker Nodes

Components of worker nodes

On worker nodes, two Kubernetes components are running—kubelet and kube-proxy.

How kubelet works

The kube-apiserver, kube-scheduler, and kube-controller are running on the control plane nodes, while kubelet runs on worker nodes where actual containers run. The kubelet nodes are the workhorses of a Kubernetes cluster. They expose computational, networking, and storage resources to containers. We can run the kubelet on bare-metal servers, virtual machines (VMs), etc.

In a nutshell, the kubelet talks to the kube-apiserver and manages the containers running on it.

The kubelet, on start-up, registers itself to the kube-apiserver by creating a dedicated Node resource. Then, the kube-scheduler can see this Node and assign new Pods running on it.

We can view all the nodes with the commands given below:

Now, let’s run the commands above in the terminal below:

The output will be as follows:

Here’s a detailed view of the nodes:

Its heartbeats are reported periodically by the kubelet. The status of the nodes can be determined by observing their STATUS column in the terminal. When a node crashes or stops reporting heartbeats for a while, it will be tainted and marked as NotReady by the ...