dharmaraj upase:
When editing kube-api server,etcd manifests files, :
• If its invalid for any reason, like volume mounts missing, invalid yaml syntax, pod dosent get created.
• how do check if yaml has issues because if its wrong pod wont get created.
• docker logs <Container_ID> for recently exiting containers shows: “Container doesnt exist.”
• sometimes, if yaml itself has issues, it wont show up in docker ps -a
Sathish Puranik:
If there are any issues with yaml files you can always look into the log file for kube api server which is present in /var/log/kube-api-* it will have exact details on what is causing for the kube-api failure…
Sathish Puranik:
You can refer more on this here https://kubernetes.io/docs/tasks/debug-application-cluster/debug-cluster/|https://kubernetes.io/docs/tasks/debug-application-cluster/debug-cluster/ for any trouble shooting your cluster.
dharmaraj upase:
we can see this by docker ps -a as well,
but that doesnt show for exited containers.
existed due to invalid yamls
Ted:
in a kubeadm environment, if the kubelet can’t start the kube-apiserver pod because of an error/problem in the yaml in the /etc/kubernetes/manifests folder, it will often show in the kubelet log
journalctl -u kubelet