🐛 Fix parsing of requests to root path of a workload cluster

[erwinvaneyk]

Summary

kcp is rejecting (valid) calls to the root path of a cluster. With a default setup, the following error is returned:

# with KUBECONFIG pointing to some workload cluster within kcp
$ kubectl get --raw='/'
[{
  "metadata": {},
  "status": "Failure",
  "message": "unable to parse cluster: no `/` found in path root",
  "reason": "BadRequest",
  "code": 400
}]

This PR resolves this issue by fixing the parsing and removing the error. With this fix:

# with KUBECONFIG pointing to some workload cluster within kcp
$ kubectl get --raw='/'
{
  "paths": [
    "/api",
    "/api/v1",
    "/apis",
    "/apis/",
    "/apis/admissionregistration.k8s.io",
    "/apis/admissionregistration.k8s.io/v1",
    "/apis/apiextensions.k8s.io",
    ...
  ]
}

[openshift-ci]

Hi @erwinvaneyk. Thanks for your PR.

I'm waiting for a kcp-dev member to verify that this patch is reasonable to test. If it is, they should reply with /ok-to-test on its own line. Until that is done, I will not automatically test new commits in this PR, but the usual testing commands by org members will still work. Regular contributors should join the org to skip this step.

Once the patch is verified, the new status will be reflected by the ok-to-test label.

I understand the commands that are listed here.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.

[ncdc]

I think it may make sense to support /, /healthz, and /readyz, but we shouldn't expose /api and /apis at the root.

[erwinvaneyk]

@ncdc Maybe I could have worded the title a bit better, but this PR solely fixes requests to the root path (/) of a workload cluster. So an endpoint like this: $HOST/apis/clusters/my-cluster/. it doesn't change anything at the root path of kcp itself

[ncdc]

Ahh ok, thanks for clarifying. I got confused from the description, because kubectl get --raw='/' ignores any path in the kubeconfig's server and literally requests /, which we don't allow. I thought you were changing that.

I do now see what your PR is doing, and it's 👍:

Before:

k --user shard-admin get --raw /clusters/root
Error from server (BadRequest): unable to parse cluster: no `/` found in path root

After:

k --user shard-admin get --raw /clusters/root
{
  "paths": [
    "/api",
    "/api/v1",
    "/apis",
    "/apis/",
    "/apis/admissionregistration.k8s.io",
...

[ncdc]

Some additional testing:

# get /clusters
Error from server (NotFound): the server could not find the requested resource

# get /clusters/
Error from server (InternalError): an error on the server ("cluster name is empty in the request context - RequestInfo: &request.RequestInfo{IsResourceRequest:false, Path:\"/\", Verb:\"get\", APIPrefix:\"\", APIGroup:\"\", APIVersion:\"\", Namespace:\"\", Resource:\"\", Subresource:\"\", Name:\"\", Parts:[]string(nil)}") has prevented the request from succeeding

I think the error for /clusters is good. Should we fix /clusters/ to also return a NotFound?

@stevekuznetsov do you want an e2e for this/these to make sure we don't regress?

[ncdc]

@erwinvaneyk would you be able to fix /clusters/ too and also add a small e2e?

[vincepri]

/lgtm

I'll follow-up with some fixes

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: ncdc

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [ncdc]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[ncdc]

/ok-to-test

[vincepri]

/retest