limitranger allow pod update

[mouuii]

What type of PR is this?

/kind feature

What this PR does / why we need it:

The LimitRanger admission plugin does not currently handle pod updates, but needs to be updated to account for InPlacePodVerticalScaling:

Which issue(s) this PR fixes:

Fixes #124855

Special notes for your reviewer:

Does this PR introduce a user-facing change?

Additional documentation e.g., KEPs (Kubernetes Enhancement Proposals), usage docs, etc.:

[k8s-ci-robot]

Hi @mouuii. Thanks for your PR.

I'm waiting for a kubernetes 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-sigs/prow repository.

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: mouuii
Once this PR has been reviewed and has the lgtm label, please assign derekwaynecarr for approval. For more information see the Kubernetes Code Review Process.

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

Needs approval from an approver in each of these files:

• plugin/pkg/admission/OWNERS

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

[mouuii]

/sig node

[mouuii]

/cc tallclair

[iholder101]

/release-note-none
/ok-to-test

Looks good overall.
Would you mind adding an e2e test?

[iholder101]

/cc

[mouuii]

/release-note-none /ok-to-test

Looks good overall. Would you mind adding an e2e test?

I'm not familiar with writing e2e and I think this change will only affect performance ？

[ffromani]

should the changes done here be gated by the InPlaceVerticalAutoscaling featuregate, like we do e.g. in the kubelet?

[iholder101]

should the changes done here be gated by the InPlaceVerticalAutoscaling featuregate, like we do e.g. in the kubelet?

Good point. +1

[mouuii]

I'm not sure if the change bring back the bug that #62673 fixed . need help .

ref #62666

@iholder101 @jennybuckley

[togettoyou]

if a.GetKind().GroupKind() == api.Kind("Pod") && a.GetOperation() == admission.Update && !utilfeature.DefaultFeatureGate.Enabled(features.InPlacePodVerticalScaling) {
	return false
}

Would it be better to still retain this part of the decision logic?

[mouuii]

The part of the decision logic is to fix the bug mentioned in #62666 . Does the bug still exist ?

The featuregate condition added just make it more confused to others . Can we find a better way to fix ?
like modify here:

kubernetes/pkg/apis/core/validation/validation.go

Lines 5250 to 5260 in fd2d352

	if !apiequality.Semantic.DeepEqual(mungedPodSpec, oldPod.Spec) {
	// This diff isn't perfect, but it's a helluva lot better an "I'm not going to tell you what the difference is".
	// TODO: Pinpoint the specific field that causes the invalid error after we have strategic merge diff
	specDiff := cmp.Diff(oldPod.Spec, mungedPodSpec)
	errs := field.Forbidden(specPath, fmt.Sprintf("pod updates may not change fields other than %s\n%v", strings.Join(updatablePodSpecFieldsNoResources, ","), specDiff))
	if utilfeature.DefaultFeatureGate.Enabled(features.InPlacePodVerticalScaling) {
	errs = field.Forbidden(specPath, fmt.Sprintf("pod updates may not change fields other than %s\n%v", strings.Join(updatablePodSpecFields, ","), specDiff))
	}
	allErrs = append(allErrs, errs)
	}
	return allErrs

BTW, unit test must modify to consider the featuregate into ,to be honest , hate the feature gate mechanism .

@togettoyou

[togettoyou]

adding a FeatureGate condition is to avoid unnecessary processing when the InPlacePodVerticalScaling feature is not enabled. As for whether the bug will recur, more test cases need to be written.

[mouuii]

I find code below allow pod to update resource request and limit , but when old pod have not set resource it will return errors ,because of the Qos .

kubernetes/pkg/apis/core/validation/validation.go

Lines 5139 to 5154 in 03ad8e5

	if qos.GetPodQOS(oldPod) != qos.ComputePodQOS(newPod) {
	allErrs = append(allErrs, field.Invalid(fldPath, newPod.Status.QOSClass, "Pod QoS is immutable"))
	}
	// handle updateable fields by munging those fields prior to deep equal comparison.
	mungedPodSpec := *newPod.Spec.DeepCopy()
	// munge spec.containers[*].image
	var newContainers []core.Container
	for ix, container := range mungedPodSpec.Containers {
	container.Image = oldPod.Spec.Containers[ix].Image // +k8s:verify-mutation:reason=clone
	// When the feature-gate is turned off, any new requests attempting to update CPU or memory
	// resource values will result in validation failure.
	if utilfeature.DefaultFeatureGate.Enabled(features.InPlacePodVerticalScaling) {
	// Resources are mutable for CPU & memory only
	// - user can now modify Resources to express new desired Resources
	mungeCpuMemResources := func(resourceList, oldResourceList core.ResourceList) core.ResourceList {

[bart0sh]

/triage accepted
/priority important-longterm