Headlamp installation on GKE giving - error loading kubeconfig files: error reading kubeconfig file

[aman-sachdeva-osttra]

Describe the bug

While installing the headlamp into Google Kubernetes engine (GKE) with WIF enabled cluster we are getting the below error:-

{"level":"error","source":"/headlamp/backend/cmd/headlamp.go","line":419,"error":"error loading kubeconfig files: error reading kubeconfig file: open /home/headlamp/.config/Headlamp/kubeconfigs/config: no such file or directory","time":"2025-03-18T12:09:05Z","message":"loading dynamic kubeconfig"

To Reproduce

Steps to reproduce the bug:

1. Create GKE cluster
2. helm repo add headlamp https://headlamp-k8s.github.io/headlamp/
3. helm install my-headlamp headlamp/headlamp
4. Check the logs of the cluster and it will give you above error

Environment (please provide info about your environment):

• Installation type: Linux
• Headlamp Version: [0.29.0 , 0.22.0]

Are you able to fix this issue?

No

Additional Context

To over come the above issue we have tried:-
1) We had changed the headlamp Dockerfile to pick the kubeconfig.yaml file from the local and place it at '/home/headlamp/.config/Headlamp/kubeconfigs/config' with the headlamp as the user. Below are the changes made:-

RUN chown -R headlamp:headlamp /headlamp
COPY headlamp-kubeconfig.yaml /home/headlamp/.config/Headlamp/kubeconfigs/config
RUN chown headlamp:headlamp /home/headlamp/.config/Headlamp/kubeconfigs/config
USER headlamp

And updated the new docker image name in the depyloyment.yaml file of headlamp, the POD was getting the config file as shown below:-

But with the above changes we were getting the same error:-

{"level":"error","source":"/headlamp/backend/cmd/headlamp.go","line":408,"error":"error loading kubeconfig files: error reading kubeconfig file: open : no such file or directory","time":"2025-03-24T11:05:04Z","message":"loading kubeconfig"}

2) We have tried with putting the kubeconfig.yaml into secrets and mounted that secrets into deployment.yaml (the mountPath was set as '/home/headlamp/.config/Headlamp/kubeconfigs/'), with passing the below env into values.yaml file which were attached to helm which doing helm install:-

env:

• name: KUBECONFIG
  value: "/home/headlamp/.config/Headlamp/kubeconfigs/config"

With the above configuration as well we are getting the same error.

Kindly help what could be the issue here, and also while going through the platforms page on the site, its states that headlamp is not sure if works with GKE or not:-

https://headlamp.dev/docs/latest/platforms#tested-kubernetes-platforms

[aman-sachdeva-osttra]

Hi Team, any input on this will be highly appreciated.

[illume]

Hello.

Let me get back to you, I think we have some documentation on this, or some other people might have done something similar.

I think it looks in .kube/config so you could try copying it there.

[SegFault02]

Hello.
Build image with kubeconfig it's 100% bad practice.

I'm use secret with kubeconfig file, and all works good.

in Deployment looks like:

volumes:
  - name: kubeconfig
     secret:
        defaultMode: 420
        secretName: headlamp-kubeconfig
spec:
  template:
    containers:
      env:
        - name: KUBECONFIG
           value: /home/headlamp/.config/Headlamp/kubeconfigs/config
      volumeMounts:
        - mountPath: /home/headlamp/.config/Headlamp/kubeconfigs/config
           name: kubeconfig
           subPath: config

You use helmchart for deploy headlamp?

[aman-sachdeva-osttra]

@SegFault02 I have tried mounting the secrets (mentioned in my 2nd point), but i was still getting the below error:-

{"level":"error","source":"/headlamp/backend/cmd/headlamp.go","line":408,"error":"error loading kubeconfig files: error reading kubeconfig file: open : no such file or directory","time":"2025-03-24T11:05:04Z","message":"loading kubeconfig"}

But when i exec into the pod the kubeconfig file was perfectly places with the correct user.

Which cloud provider you are working with, is it GCP?