Kubernetes 101 : Static pods, their drawbacks and the kubelet daemon
We can deploy "unmanaged" static pods on a specific host, of our choosing, by running the below command on that node:
Then we check if the pod is running, using the below command:
We can stop all the pods running on the "worker-node", by preparing it for "maintenance" mode, using the below command:
"kubectl drain" doesn't stop the pods that for some reason can't be stopped, we could use the "--force" parameter to override that behavior.
The "nginx-pod" on the "worker-node" is managed by the kubelet that's running on that specific node.
After "draining" the node, we try to display the pods on our worker node as below, but we don't get any output:
To prepare our node for "housing" pods again, we could then restart the kubelet daemon as below:
And to make our node available for scheduling pods again, we use the below command: