[FG:InPlacePodVerticalScaling] Pod Resize - resize stuck "InProgress" when only resizing memory requests about kubernetes

[hshiina]

It looks that nothing is requested to the runtime when only a memory request is changed:

kubernetes/pkg/kubelet/kuberuntime/kuberuntime_manager.go

Lines 589 to 590 in 5df8e15

	// Note: cgroup doesn't support memory request today, so we don't compare that. If canAdmitPod called during
	// handlePodResourcesResize finds 'fit', then desiredMemoryRequest == currentMemoryRequest.

Even if a resizePolicy is RestartContainer, a container is not restarted in this case. requests.memory in a container status is copied from allocatedResources:

kubernetes/pkg/kubelet/kubelet_pods.go

Lines 2112 to 2114 in 5df8e15

	if memory, found := status.AllocatedResources[v1.ResourceMemory]; found {
	requests[v1.ResourceMemory] = memory.DeepCopy()
	}

I think the problem in this issue is that it takes 60-90 seconds to update requests.memory in a container status. Currently, this is updated at a regular resync for the pod because nothing triggers status update. It would be better to update it earlier.

[k8s-ci-robot]

This issue is currently awaiting triage.

If a SIG or subproject determines this is a relevant issue, they will accept it by applying the triage/accepted label and provide further guidance.

The triage/accepted label can be added by org members by writing /triage accepted in a comment.

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.

[dshebib]

/sig node

[hshiina]

/assign

[hshiina]

Could you please confirm if the requests.memory in the container status is updated to 400Mi 90 seconds after the resize request is patched?