Update pod resize test to accept new cpu.weight conversion.

[bitoku]

What type of PR is this?

/kind failing-test

What this PR does / why we need it:

Because of the changes in conversion of cpu.weight in cgroupv2, the InPlacePodVerticalScaling tests no longer pass.

• cgroup: change conversion from CPU shares to weight containers/crun#1767
• ConvertCPUSharesToCgroupV2Value: improve opencontainers/cgroups#20

This change enables graceful migration from the old conversion to the new conversion, not to break the CI.

Which issue(s) this PR is related to:

• Conversion of cgroup v1 CPU shares to v2 CPU weight causes workloads to have low CPU priority #131216

Special notes for your reviewer:

Does this PR introduce a user-facing change?

None

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

[k8s-ci-robot]

Hi @bitoku. 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.

[haircommander]

can you add a link to the oci runtime changes/issues to track why we check both?

[bitoku]

I added the description. let me know if anything unclear.

[giuseppe]

LGTM

[giuseppe]

There are other instances of this conversion in the code. Will they be addressed separately?

[bitoku]

@giuseppe I just didn't realize other instances. I'd like to include them in this PR. Can you point out for me?

[giuseppe]

@giuseppe I just didn't realize other instances. I'd like to include them in this PR. Can you point out for me?

these are the ones I see:

➜ git grep 262142
pkg/kubelet/cm/cgroup_manager_linux.go: return 1 + ((*cpuShares-2)*9999)/262142
pkg/kubelet/cm/cgroup_v2_manager_linux.go:      return uint64((((cpuShares - 2) * 9999) / 262142) + 1)
pkg/kubelet/cm/cgroup_v2_manager_linux.go:      return uint64((((cpuWeight - 1) * 262142) / 9999) + 2)
test/e2e/common/node/framework/cgroups/cgroups.go:              return 1 + ((shares-2)*9999)/262142
test/e2e_node/node_container_manager_test.go:   return 1 + ((shares-2)*9999)/262142

most of them are mechanical, except return uint64((((cpuWeight - 1) * 262142) / 9999) + 2) that needs the inverse function

[k8s-ci-robot]

LGTM label has been added.

Git tree hash: 30fca49d6f5e49a116ee1a3ee78a4a3284186a89

[soltysh]

@bitoku one nit, can you split the commits, such that one is touching go.mod, go.sum and vendor/ directory, and the other one is where you're modifying the test code?

[bitoku]

Hi @soltysh, thanks for the review.
Could you tell me the reason for splitting?

[soltysh]

Hi @soltysh, thanks for the review. Could you tell me the reason for splitting?

It's easier to review, since the actual code changes are handled manually and require more thorough checks, whereas the automatic vendor changes less so. It's a general practice used in the project to have these two separate. Something similar applies to changes coming from running make update. Although, I will say it's not strictly required 😉

[sohankunkerkar]

Is there a plan to backport this to 1.33 once it’s merged?

[bitoku]

@sohankunkerkar
I don't think we need to backport it.
It changes only the test. Unless we update crun/runc, the test in 1.33 (and 1.34 too) won't break.

[SergeyKanzhelev]

the name implies that this file will have some build directives

[bitoku]

The file suffix is also respected as a build constraint.
https://pkg.go.dev/go/build#hdr-Build_Constraints

These files don't have one either.

kubernetes/pkg/kubelet/cm/util/cgroups_linux.go

Lines 1 to 3 in 5ecddb6

	/*
	Copyright 2016 The Kubernetes Authors.

kubernetes/cmd/kubelet/app/server_linux.go

Lines 1 to 3 in 7ddb042

	/*
	Copyright 2015 The Kubernetes Authors.

[SergeyKanzhelev]

beyond the build directive this LGTM

[SergeyKanzhelev]

/lgtm
/approve

[bitoku]

@liggitt Can you PTAL as dep-approver?

[liggitt]

was it intentional to change the serialization upstream? this will now serialize explicit null values... does that impact us in any user-facing way?

[haircommander]

I don't believe the kubelet serializes this cgroup object at all, it's kept as a go structure AFAIU and not read to or from anything. It's possible CRI implementations do, but this vendor doesn't affect that... we only use it in libctCgroupConfig and in every case it's used it's immediately passed to libctr.New() function

[liggitt]

/approve
for dep changes

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: bitoku, giuseppe, liggitt, SergeyKanzhelev

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 [liggitt]
• test/e2e/common/OWNERS [SergeyKanzhelev,liggitt]
• vendor/OWNERS [liggitt]

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

[k8s-triage-robot]

The Kubernetes project has merge-blocking tests that are currently too flaky to consistently pass.

This bot retests PRs for certain kubernetes repos according to the following rules:

• The PR does have any do-not-merge/* labels
• The PR does not have the needs-ok-to-test label
• The PR is mergeable (does not have a needs-rebase label)
• The PR is approved (has cncf-cla: yes, lgtm, approved labels)
• The PR is failing tests required for merge

You can:

• Review the full test history for this PR
• Prevent this bot from retesting with /lgtm cancel or /hold
• Help make our tests less flaky by following our Flaky Tests Guide

/retest