Re: [kubernetes/kubernetes] "kubectl explain" should be able to explain "apiservices" and "customresourcedefinition" (#49465)

[k8s-ci-robot]

@xiangpengzhao: Reiterating the mentions to trigger a notification:
@kubernetes/sig-cli-bugs.

In response to this:

@kubernetes/sig-cli-bugs

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.

—
You are receiving this because you are on a team that was mentioned.
Reply to this email directly, view it on GitHub, or mute the thread.

[Chao Xu]

cc @mbohlool

[Mehdy Bohlool]

This look like a problem with old swagger 1.2. We do not aggregate swagger 1.2 spec. When kubectl move to OpenAPI, it should by fix this by design.

[Kubernetes Submit Queue]

Closed #49465 via #53228.

[k8s-ci-robot]

@nikhita: you can't re-open an issue/PR unless you authored it or you are assigned to it.

In response to this:

This is still not fixed (even after kubectl moving to OpenAPI).

There are two problems here:

1. kubectl uses groups registered into a group registry. This can be fixed by:

--- a/pkg/kubectl/cmd/explain.go
+++ b/pkg/kubectl/cmd/explain.go
@@ -26,7 +26,6 @@ import (
	"k8s.io/kubernetes/pkg/kubectl/cmd/templates"
	cmdutil "k8s.io/kubernetes/pkg/kubectl/cmd/util"
	"k8s.io/kubernetes/pkg/kubectl/explain"
-	"k8s.io/kubernetes/pkg/kubectl/scheme"
	"k8s.io/kubernetes/pkg/kubectl/util/i18n"
)

@@ -91,7 +90,6 @@ func RunExplain(f cmdutil.Factory, out, cmdErr io.Writer, cmd *cobra.Command, ar
		return err
	}

-	// TODO: We should deduce the group for a resource by discovering the supported resources at server.
	fullySpecifiedGVR, groupResource := schema.ParseResourceArg(inModel)
	gvk := schema.GroupVersionKind{}
	if fullySpecifiedGVR != nil {
@@ -104,20 +102,12 @@ func RunExplain(f cmdutil.Factory, out, cmdErr io.Writer, cmd *cobra.Command, ar
		}
	}

-	if len(apiVersionString) == 0 {
-		groupMeta, err := scheme.Registry.Group(gvk.Group)
-		if err != nil {
-			return err
-		}
-		apiVersion = groupMeta.GroupVersion
-
-	} else {
-		apiVersion, err = schema.ParseGroupVersion(apiVersionString)
-		if err != nil {
+	if len(apiVersionString) != 0 {
+		if apiVersion, err = schema.ParseGroupVersion(apiVersionString); err != nil {
			return err
		}
+		gvk = apiVersion.WithKind(gvk.Kind)
	}
-	gvk = apiVersion.WithKind(gvk.Kind)

	resources, err := f.OpenAPISchema()
	if err != nil {

However, after this it gives:

error: Couldn't find resource for "apiextensions.k8s.io/v1beta1, Kind=CustomResourceDefinition"

2. This occurs because of incorrect openAPI schema. The schema does not contain apiextensions and apiregisteration.

kubectl explain looks up the gvk of the resource in question from a list of resources:

https://github.com/kubernetes/kubernetes/blob/ff626a35fe35f6d219a3f95b39df7afd4cab9b6b/pkg/kubectl/cmd/util/openapi/openapi.go#L71-L77

It gets the gvk here:

https://github.com/kubernetes/kubernetes/blob/ff626a35fe35f6d219a3f95b39df7afd4cab9b6b/pkg/kubectl/cmd/util/openapi/openapi.go#L59

parseGroupVersionKind finds the gvk through groupVersionKindExtensionKey:

https://github.com/kubernetes/kubernetes/blob/ff626a35fe35f6d219a3f95b39df7afd4cab9b6b/pkg/kubectl/cmd/util/openapi/openapi.go#L83-L87

Here, groupVersionKindExtensionKey is "x-kubernetes-group-version-kind" and is the key used to lookup the GroupVersionKind value for an object definition from the definition's "extensions" map.

However, we do not generate x-kubernetes- extensions for apiextensions-apiserver and aggregator yet (#52741), due to which it fails.

/reopen

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.

—

[Nikhita Raghunath]

—

[Peter (XiangPeng) Zhao]

[k8s-ci-robot]

Reopened #49465.

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[Nikhita Raghunath]

/remove-lifecycle stale

[Nikhita Raghunath]

Looks like the blocker was fixed in #64174 🎉

I will move forward with the fix.

[Nikhita Raghunath]

/assign

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

—

[Nikhita Raghunath]

/remove-lifecycle stale

[Jordan Liggitt]

Closed #49465.

[Jordan Liggitt]

fixed in 1.11 in #64174