Use newer v1beta1 version in API version upgrade test

[CecileRobertMichon]

What type of PR is this?

/kind failing-test

What this PR does / why we need it: The k8s 1.22 images have been deprecated and removed from the Azure marketplace. Since CAPI v1.0.x only supports old k8s versions, we are no longer able to run the test as-is. This PR changes the upgrade test to run upgrade from the earliest supported minor release of CAPI and CAPZ (n-2).

Which issue(s) this PR fixes (optional, in fixes #<issue number>(, fixes #<issue_number>, ...) format, will close the issue(s) when PR gets merged):
Fixes #4422

Special notes for your reviewer:

• cherry-pick candidate

TODOs:

• squashed commits
• includes documentation
• adds unit tests

Release note:

NONE

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (285a75b) 61.96% compared to head (25b53a8) 61.96%.
Report is 8 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #4433   +/-   ##
=======================================
  Coverage   61.96%   61.96%           
=======================================
  Files         188      188           
  Lines       18768    18768           
=======================================
  Hits        11630    11630           
  Misses       6500     6500           
  Partials      638      638           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[nojnhuh]

Overall LGTM.

Should we also document in the release guide to make these same changes whenever we cut a new minor release?

[nojnhuh]

Should we also document in the release guide to make these same changes whenever we cut a new minor release?

I think even a link to this PR would be good enough.

[CecileRobertMichon]

/hold

[nojnhuh]

/lgtm

[k8s-ci-robot]

LGTM label has been added.

Git tree hash: 900689e6eb81f7cad939369e688cd9b68ed8f355

[nojnhuh]

/assign @willie-yao

[willie-yao]

Looks good pending one question!

[CecileRobertMichon]

/retest

[willie-yao]

/lgtm

[nojnhuh]

/approve

@CecileRobertMichon please remove the hold whenever you're ready.

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: nojnhuh

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 [nojnhuh]

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

[mboersma]

/lgtm++

Thanks @CecileRobertMichon!

[nojnhuh]

@CecileRobertMichon Do we need to cherry pick this to fix the test on the release branch?

[CecileRobertMichon]

Do we need to cherry pick this to fix the test on the release branch?

Yes. How do we want to handle the upgrade from version on previous release branches? Should it always be n-2? In which case we might need a manual cherry-pick.

[CecileRobertMichon]

/hold cancel

[nawazkh]

/lgtm

[nawazkh]

Is there a way we could fetch these values from azure-dev.yaml so that we do not maintain it going ahead? Not a blocking comment per se.

[CecileRobertMichon]

since we're adding a second test with a different set of versions we need to configure this per test #4446

We could have two separate sets of env variables 🤔 Maybe something to follow up on in #4446 @willie-yao? Let's merge this one as-is for now to fix the broken test

[willie-yao]

I agree that having two sets of env variables would be the better way to go about this. Will work on implementing it for my PR!