Server-Side Apply status wiping

[kevindelgado]

What type of PR is this?

/kind bug

What this PR does / why we need it:

Implements kubernetes/enhancements#1123
Adds the ResetFieldsProvider interface and implements it for all strategies.

Which issue(s) this PR fixes:

Fixes #75564

Does this PR introduce a user-facing change?

FieldManager no longer owns fields that get reset before the object is persisted (e.g. "status wiping").

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

- [KEP]: https://github.com/kubernetes/enhancements/pull/1123

[k8s-ci-robot]

Hi @kevindelgado. 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/test-infra repository.

[kevindelgado]

/assign @liggitt

We need to fix one broken test case in status_test, but wanted to get your eyes on this in the meantime @liggitt to begin reviewing this approach to status wiping.

The first commit is from @kwiesmueller work from back in July #92809
and we've recently added some fixes to broken test cases (second commit in the PR onwards).

cc @apelisse

[apelisse]

/ok-to-test

[kevindelgado]

/test pull-kubernetes-e2e-kind

[kevindelgado]

/test pull-kubernetes-integration

[kevindelgado]

Does this need to have a 1.21 milestone label in order to land in the release @apelisse?

[apelisse]

I don't think it does, but it wouldn't hurt

[apelisse]

Looks good, awesome work to both Kevin :-)
/lgtm

[lavalamp]

/milestone v1.21
/lgtm

[lavalamp]

/lgtm

[apelisse]

/lgtm

[kwiesmueller]

Thanks for finishing the job!

[lavalamp]

I manually removed the "needs-rebase" label, I'm not sure why the bot didn't.

[liggitt]

This is passing in newObj for both args, which means ResetObjectMetaForStatus is a no-op.

Also, this previously allowed updating finalizers via status updates, and ResetObjectMetaForStatus does not allow that. The crd finalizer depends on updating finalizers via a status update API call, so simply fixing this to pass newObj/oldObj correctly will break the CRD finalizer controller.