Browser user/password? about skooner

[herbrandson]

Thank you for helping track it down :)

Deploying the fix now. Should be available on the k8dash:latest tag from docker hub in the next 10-15 minutes.

[herbrandson]

Great. Just waiting for the build to clear the docker hub queue. I'll keep you posted. Thanks again!

[jdn-za]

Just got back to this, thanks for the fix

[boosis]

I did exactly the same but instead of user and password, it asked me to paste the token. And once I did that, I was able to login.

[zetaab]

I am using OIDC for login, i need to press like two times cancel to user+password and then it will login and works

[herbrandson]

I wasn't able to reproduce this last evening. I thought that perhaps it was because the cluster I was testing against was using OIDC, but based on the comment from @zetaab it sounds like the issue was happening even with OIDC. Can you provide any further information about your cluster? How was it installed (i.e. kubeadm, kops, eks, gke)?

[scottconrad88]

@herbrandson I was getting the same issue as @jdn-za above when I was using kubectl port-forward ... I switch it over to use the Ing resource you mentioned in the README but saw the same error.

I dug into the nginx logs and found this GET request

172.20.37.44 - [172.20.37.44] - - [02/Apr/2019:15:52:12 +0000] "GET /oidc HTTP/1.1" 200 2

I don't have OIDC enabled on the cluster I wanted to test this out on but looks like it might be defaulting to it..?

[herbrandson]

Thanks for the feedback @scottconrad88. That api is actually expected even when OIDC is not enabled. That call is actually checking with the backend to see if OIDC is configured or not.

It's interesting that you see it even using the ingress controller. Is there any chance you could open the browser developer tools to see if we can figure out which network call is generating the prompt?

Also, what version of Kubernetes are you using?

[scottconrad88]

Ah sorry about that @herbrandson didn't dive into that endpoint to much. That makes sense. The endpoint that is generating the request is Request URL: https://dash.containers.rocketrip.com/openapi/v2 Below is a screen shot and what the network request looks like.

I am on Kubernetes version 1.10.11 and fyi I have the metrics server installed.

[herbrandson]

Thanks @scottconrad88. That's really helpful. I think I know what's causing this now. I've pushed a fix to the dev branch. It's currently building a new container on docker hub under the k8dash:dev label. Any chance someone would be will to test it out?

[scottconrad88]

No problem @herbrandson. I'll update my k8s yaml file to use that image and I can test it out.

[herbrandson]

Ok, it's built and ready to go. Just change the image in the yaml to herbrandson/k8dash:dev. Let me know how it goes.

[scottconrad88]

@herbrandson yup looks like that fixed it. Thanks for the quick turnaround!