Join GitHub today
"the server has asked for the client to provide credentials" #277
When using a kubeconfig that contains many clusters, contexts and users, if the current context does not match the current env, ksonnet returns
What you expected to happen:
Applying ksonnet envs shouldn't rely on the context being set with kubectl
How to reproduce it (as minimally and precisely as possible):
Anything else we need to know?:
This may be just a misunderstanding on my part, but if ks has the ability to know about servers, namespaces and other things, when creating an env, it should have enough information to apply the changes. Perhaps if ks stored the user as well, it would be more useful.
The reason I think this is important is because in a CI pipeline, it would be useful to not have to specify
@jaxxstorm Thanks for the report, that should not be the case! I can't seem to reproduce it, but I suspect that it's trying to authenticate with incorrect user credentials, so: are you using different users in X and Y? If not, could you share X and Y (without the key information, of course :) )?
Here's my kubeconfig:
In this case, X is uw2d1 and Y is uw2d2. If there's anything else I can provide, let me know
@jaxxstorm Ok, so what's happening is that
I believe this is a fairly common case, so we will need to design some kind of a workflow for how to bind users to different environments.
@jaxxstorm I'd like to review what we have here. If I understand correctly, it looks like you are having issues with ksonnet determining the correct context to run with. As it stands right now, we store the cluster's API server URL and a namespace. This is not optimal. In a better world, we should expose multiple ways of identifying a cluster (including specifying a context). This task isn't trivial, but should be straight forward.
In the mean time, does specifying the
The next steps are to: