Deekshith Hadil:
ah… this environment is setup in different way. Kubelet is probably running as some container.
Are you getting anything while running ps -eFwww | grep kubelet
Sampathkaran:
is this a kubeadm setup ?> are you using the lab instance ?
Pradeep.nair:
Yes lab instance … Practices test for control plane failure
Sampathkaran:
i believe some issue with the lab not loaded properly …try to create a new lab instance and see. I never seen such issues when i m doing my labs
Pradeep.nair:
is der any way to check the below kubeadm.conf file i wanted to check the below path
–kubelet_config_args=–config=<path ofthe kubelet config.yaml>
Deekshith Hadil:
Deekshith Hadil:
Definitely this is configured in different way. Not kubeadm or hardway.
I would encourage you to do some local setup and find it by yourself
Which lab practice is this? I would like to have a look around.
Pradeep.nair:
@Sampathkaran i try to reload this again and again same isssue
@Deekshith Hadil this is lab for Practices test control plane failure
Pradeep.nair:
i mean only the location is missing to check the path… we can go and do the labs no issue … but can;t see the config file set in kubeadm.conf file …
Deekshith Hadil:
hold on. let me try this lab
Deekshith Hadil:
@Pradeep.nair
Is this 239. Practice Test - Worker Node failure?
Deekshith Hadil:
I looked at
> 236. Practice Test - Control Plane Failure
> 239. Practice Test - Worker Node Failure
The kubelet related questions were in 239.
The issue was on the node node01.
The screenshots you shared shows that you are checking on controlplane. All the questions had issues on node node01.
Pradeep.nair:
@Deekshith Hadil i was doing 236. Practice Test - Control Plane Failure .it is first question … where kube-controller-manger is in crash state and need to check the file in /etc/kubernetes/manifest location … the issue was in one of pod -n kube-system namespace which resides in control plane not node01
Deekshith Hadil:
ah…
Deekshith Hadil:
It is not problem with kubelet. You should be looking at kube-controller-manager.
it is running as statick pod.
So
$ kubectl get pods -n kube-system
which shows some issues with kube-controller-manager pod.
you can try.
$ kubectl describe pod kue-controllxxxxxxxx -n kube-system
To look at the events.
If still not found anything,
$ kubectl logs <pod-name> -n kube-system
or
$ docker container ls --all | grep controller
$ docker container logs <container_id>
This will take you to the error.
Pradeep.nair:
Will tried the lab again to find the docker events
Deekshith Hadil:
I just tried again. It has nothing to do with kubelet.
ns680:
I also have seen this issue which confused me a while back. I came to the conclusion that Kodecloud clusters are simulated not real and its nodes are run as containers on top of a cloud provider. They are great if you stick to the course lab material.