kubeadm: write "cgroupfs" in the kubelet config

[neolit123]

Reason for PR:

Cgroup drivers should be ignored on Windows but there is a bug
in the kubelet:
kubernetes/kubernetes#97759
the bug is fixed in master (.21) but is not part of the latest stable releases in the skew yet.

Workaround this bug by patching the kubelet configuration in
the cluster.

Issue Fixed:

fixes #138

Requirements

• Sqaush commits
• Documentation
• Tests

Notes:

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by: neolit123
To complete the pull request process, please assign jsturtevant after the PR has been reviewed.
You can assign the PR to them by writing /assign @jsturtevant in a comment when ready.

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

Needs approval from an approver in each of these files:

• OWNERS

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

[neolit123]

/assign @jsturtevant

[neolit123]

the bug is fixed in master (.21) but is not part of the latest stable releases in the skew yet.

we should have backports for this:
kubernetes/kubernetes#97764 (comment)

[neolit123]

/kind failing-test
/priority critical-urgent

[jsturtevant]

I am not super familiar with this setup. It looks like the cherrypicks went in for older versions: kubernetes/kubernetes#98384

Should we prefer those over the patch?

[neolit123]

@jsturtevant

i've sent the PR after investigating the failures and so that we don't delete the job due to not being maintained.
cc @jayunit100

these scripts were originally created by benmoss with some help from me, but ideally sig-windows should take ownership.
your observation about the cherry picks is correct - once a new k8s PATCH release is out for the latest stable (1.20) this setup should no longer fail and this patch will not be needed.

we missed v1.20.2 for the cherry pick, so it will be out with v1.20.3, maybe in ~2 weeks.
/hold

[jsturtevant]

Sounds good. FWIW we have additional coverage for windows Kubeadm over in capz: https://testgrid.k8s.io/sig-cluster-lifecycle-cluster-api-provider-azure#periodic-capi-e2e-full

It's not as extensive (I am working on that) but it does validate that kubeadm binary works. I'll need to make sure we get the correct fix there too.

[neolit123]

given capz is more actively maintained than the CI scripts in this repository, we can remove them and if that is what sig-windows would prefer.

[neolit123]

1.20.4 was just released:
https://groups.google.com/d/msgid/kubernetes-dev/pYK9msu4TUqsFh7mUQiCXQ%40geopod-ismtpd-1-0?utm_medium=email&utm_source=footer

will check if the latest kubelet there fixes the related failures:
https://k8s-testgrid.appspot.com/sig-windows-gce#kubeadm-windows-gcp-k8s-stable

[neolit123]

the job deployment is now working:
https://k8s-testgrid.appspot.com/sig-windows-releases#kubeadm-windows-gcp-k8s-stable

some conformance test appear to be red, which is not deployer related.

/close

[k8s-ci-robot]

@neolit123: Closed this PR.

In response to this:

the job deployment is now working:
https://k8s-testgrid.appspot.com/sig-windows-releases#kubeadm-windows-gcp-k8s-stable

some conformance test appear to be red, which is not deployer related.

/close

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.