Have three hosts to run Rancher cluster.
Install k8s from catalog on master host.
Set orchestration=true
and etcd=true
labels to two Rancher agent hosts.
After the k8s stack finished, only the kubelet
went wrong. Unhealthy
with 0 containers.
Why?
The question has been debugged in the comment section.
Kubernetes Mantra
I have added some additional point to keep it in mind to debug the Kubelet.
The K8s cluster is made of Masters and Workers Node which has several components. Kubelet is one the component which needs to take care properly.
Let's begin by saying that Master node manages or orchestrate the cluster state and Workers node run the pods.However, Without Kubelet It does not work Since It will be part of each node whether it's a Master or Worker. Performance of the cluster certainly depends on the kubelet.
We can use the following command to check its status and activity or logs.As It is deployed as system-service by systemd.
systemctl status kubelet
journalctl -xeu kubele