-
Notifications
You must be signed in to change notification settings - Fork 4.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Need help configuring dashboard correctly #628
Comments
After some research, I think the answer might be:
Is this accurate? Did I miss anything? |
This is all correct. The code we use is kubernetes client package located at: https://github.com/kubernetes/kubernetes/blob/dae5ac482861382e18b1e7b2943b1b7f333c6a2a/pkg/client/unversioned/clientcmd/doc.go Can you help us adding this to the documentation? Or fixing the initialization phase to automatically start on your setup? Do other addons, like Heapster, work out of the box? |
I'm fixing my setup. I think it'd be useful to document as well, so other people can figure it out more easily. I'm happy to write some docs on it: where would you suggest they go, exactly? |
Awesome, thank you! In this repository it is best to put is in |
I wrote some docs. I would have written something with a larger scope, but I don't yet know enough. I'm happy to tweak this in any way that you want, so it fits into the existing docs well. I hope this helps. |
I'm developing a new kube-up implementation for VMware's Photon Controller. While the basics work fine (I can deploy pods and services), I can't get the UI to work. I have the same symptom as in Issue #374. That is, when I connect to the UI through my web browser, I get an error that it can't connect to localhost:8080. If I pass --apiserver-host, I get an error like:
Get https://10.146.37.191/api/v1/replicationcontrollers: x509: failed to load system roots and no roots provided
I'm experiencing this with both the 1.2 release of Kubernetes as well as the head of the master branch. I've used both dashboard 1.0.0 and 1.0.1
Unfortunately, a combination of studying issue #374 and lots of experimentation hasn't helped. (I've mounted /etc/ssl/certs and ensured a .kube/config file was in place, among other things.) I'm sure the fault is all mine: I must not have the configuration set up correctly.
Instead of asking you "what's wrong?", I'd prefer to ask you how the dashboard knows how to connect to the API server. How does it normally discover the IP address and the correct way to authenticate?
I've noticed that the dashboard container doesn't mount any of the standard places to find a kubeconfig file, so it must be getting the configuration another way. Is it via the environment? What environment variables is it looking for? I'm guessing it might use
KUBERNETES_MASTER
to find the IP address, but I don't know how it finds the authentication configuration. I've looked at the Kubernetes client code a bit, but it wasn't immediately obvious.Like many of the kube-up implementations--but unlike the AWS and GCE implementations--my Kubernetes nodes's kubeconfig files for kubelet and kube-proxy use tokens instead of certificates. Could the fact that I haven't set up certificates be a problem? If I do set them up, where does the dashboard expect to find them?
Thanks!
The text was updated successfully, but these errors were encountered: