e2e: allow unknown providers with a warning

[pohly]

What type of PR is this?
/kind bug

What this PR does / why we need it:

#68483 unintentionally broke several use cases by tightening the validation of the --provider parameter such that the e2e.test binary only accepts known providers.

Whether that really is the right behavior needs more discussion (and in hindsight it probably isn't). In the meantime we need to restore the previous behavior quickly to get all CI jobs working again.

Which issue(s) this PR fixes (optional, in fixes #<issue number>(, fixes #<issue_number>, ...) format, will close the issue(s) when PR gets merged):
Refs #70058

Special notes for your reviewer:

Does this PR introduce a user-facing change?:

- Support for passing unknown provider names to the E2E test binaries is going to be deprecated. Use `--provider=skeleton` (no ssh access) or `--provider=local` (local cluster with ssh) instead.

/sig testing
/cc @BenTheElder
/cc @neolit123

[neolit123]

LGTM thanks.
maybe @timothysc has historic knowledge of why we allowed unknown providers pre-refactor.
/assign @timothysc

[neolit123]

x reference this PR which is on hold with the alternative solution:
kubernetes/test-infra#9897

this tackles the problem in kubetest.

if this current PR is merged we need to cancel the above PR..
also we need to revert this test-infra PR:
kubernetes/test-infra#9884

/this-is-fine

[k8s-ci-robot]

@neolit123:

In response to this:

x reference this PR which is on hold with the alternative solution:
kubernetes/test-infra#9897

this tackles the problem in kubetest.

if this current PR is merged we need to cancel the above PR..
also we need to revert this test-infra PR:
kubernetes/test-infra#9884

/this-is-fine

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.

[timothysc]

So your change exposed s bunch of brittleness in the tests, this is kind of a band-aide, but do we still need it? Are the other tests fixed?

/hold

[pohly]

@timothysc At the moment this band-aid is still needed. I prefer to have it in 1.13 with a deprecation announcement, then in 1.14 remove it (and enhance the error handling - Failf produces rather ugly output). I have updated this PR's release note accordingly, in case that it gets merged.

[timothysc]

@pohly please log an issue and add a // TODO comment to the top of the conversion referencing the issue(s), and I'll get this unblocked.

[BenTheElder]

Discussed at the kubeadm office hours, one upshot to this is that removing providers from core should be slightly easier since they just get promoted to no-op with a warning.

We're also fixing tests to avoid this warning.

[pohly]

@timothysc I've file an issue and added a TODO in the code - I hope I understood correctly where you wanted that.

[timothysc]

/approve
/hold cancel

@BenTheElder for LGTM b/c he knows the deets of how it's currently failing.

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: pohly, timothysc

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:

• test/OWNERS [timothysc]

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

[BenTheElder]

/lgtm
thanks!