Hi on the 'Troubleshooting section' the test 'Practice Test -Troubleshoot Networ . . .

Kubernetes
#1

Sandy Mauriz:
Hi on the ‘Troubleshooting section’ the test ‘Practice Test -Troubleshoot Network’ got some issues on question 2:
After changing the kube-proxy daemonset from: /var/lib/kube-proxy/config to /var/lib/kube-proxy/kubeconfig.conf
I get a different error from the logs:

1 server.go:490] failed complete: failed to decode: no kind “Config” is registered for version “v1” in scheme “http://k8s.io/kubernetes/pkg/proxy/apis/config/scheme/scheme.go:29|k8s.io/kubernetes/pkg/proxy/apis/config/scheme/scheme.go:29”

And when going to question 3 the issue still doesn’t get resolved so I cannot do it either.

Could you please help, thanks

#2

Tej_Singh_Rana:
Hello, @Sandy Mauriz @Vic
Please check again kube-proxy configMap. Correct file name is config.conf instead of kubeconfig.conf .

#3

Mohamed Ayman:
Check this solution:
NetworkTroubleshooting-Q2.gif

#4

Sandy Mauriz:
Hi @Mohamed Ayman @Tej_Singh_Rana thanks for that, but still after modifying the ds the kube-proxy is still on error but a different one. The new error it says ‘no kind Config’

#5

I still face the same error as well

please advise

controlplane $ kubectl logs -n kube-system kube-proxy-g675h 
F0514 20:21:36.897793       1 server.go:490] failed complete: failed to decode: no kind "Config" is registered for version "v1" in scheme "k8s.io/kubernetes/pkg/proxy/apis/config/scheme/scheme.go:29"

please find below a similar issue where the supplied config file was not relevant to the kube-proxy

#6

Hi all

I repeated the exercise and checked the parameters and apparently the below parameter was configured in the daemon set which is mentioned in the answer guide

/var/answers/answer2.md 



  --config=/var/lib/kube-proxy/config.conf

so I assume that you did the same mistake I did and used the one mentioned in the configMap.

/var/lib/kube-proxy/kubeconfig.conf
2 Likes
#7

Hi all

i just did exactly as @noursh , found [ /var/lib/kube-proxy/kubeconfig.conf ] in the configmap and did use it to fix the DaemonSet, wich is not the correct to do.

the definition of the file name is at the begin of the datamap definition (between annotation and apiversion) and look like bellow. – to get it run [ kubectl -n kube-system describe configmaps kube-proxy ]

CONFIGMAP DEFINITION

##################################################
Annotations:  kubeadm.kubernetes.io/component-config.hash: sha256:c15650807a67e3988e859d6c4e9d56e3a39f279034149529187be619e5647ea0

Data
====
config.conf:
----
apiVersion: kubeproxy.config.k8s.io/v1alpha1
##################################################

So, go ahead to fix your DaemonSet to point to [ /var/lib/kube-proxy/config.conf ] and boom! Get it done!

Thank you @noursh , your post did clarify a two hours investigation!