Set up minikube for k8s testing

[umohnani8]

Install and set up minikube so that we can
create a k8s cluster for testing.

Signed-off-by: Urvashi Mohnani [email protected]

Does this PR introduce a user-facing change?

None

[umohnani8]

@cevich PTAL

[cevich]

Looks like a good start to me. To speed up development, please feel free to do a temporary DO NOT MERGE commit that removes the rootless_system_test dependency. For other tasks (except build), you can bypass them by (temporarily) adding a skip: $CI == $CI item to them. That'll should get the minikube task running earlier so you can iterate faster 😁

[cevich]

Oh, and you'll want to change IMAGE_SUFFIX: "c5823947156488192" to IMAGE_SUFFIX: "c5167094042984448". The older images don't have the minikube rpm cached in them.

[cevich]

FYI: The new task is throwing Error: Unable to find a match: /var/cache/download/minikube-latest* because the IMAGE_SUFFIX in .cirrus.yml is wrong (see my comment above).

[umohnani8]

@cevich are these failures related to my changes? Did I mess up something?

[umohnani8]

@cevich any idea why the minikube test was cancelled?

[edsantiago]

@umohnani8 the minikube test depends on rootless_system_test, which in turn depends on rootless_integration_test, which failed. The failure looks real, because many tests failed in the same way. I suspect a problem with the VM images.

[cevich]

Ahh this is the "os error 95" thing. This is caused by a bug in netavark (I think). There's a fix that should be in the stable repos by now.. Getting it into CI means building new VM Images. I've got an unrelated PR doing that now, I just need to flip some switches. Can you wait a day or two or you need something sooner?

[umohnani8]

@cevich thanks, I can wait a day or so for it to be done.

[cevich]

Great, if that changes let me know and I'll do a separate build for this. Otherwise they can come from containers/automation_images#213 - if only the damn CI would pass (lol...it's my fault, I'm trying to add new tests).

[cevich]

okay @umohnani8 I think this should work. Give IMAGE_SUFFIX: "c4678746211876864" a try and let me know.

[TomSweeneyRedHat]

LGTM
once the tests are hip. Hopefully @cevich 's thought cures all ills, at least in this PR.

[edsantiago]

Came here to agree with @cevich's assessments, found a few more concerns

[edsantiago]

This is moving waaaaaaaaaaaaay too fast for slow old me. I will step back until things calm down (like, tomorrow morning).

[edsantiago]

Final set of comments on this iteration. All of them are nits: in broad sense, looks like good testing. Thanks for your tolerance of my nitpicking!

[edsantiago]

Reviewing again. The new helpers.bash needs a trailing newline to pass validation. Since you need to resubmit anyway, would you consider calling basic_setup from your setup? It will prefetch $IMAGE and do some cleanup that might give you an extra edge in running tests after one test fails.

[umohnani8]

Thanks for all the reviews @edsantiago and @cevich!

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: edsantiago, umohnani8

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 [edsantiago,umohnani8]

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

[edsantiago]

New tests failing:

[+0004s] not ok 1 minikube - check cluster is up
         # (from function `setup' in test file test/minikube/[helpers.bash, line 8](https://github.com/containers/podman/blob/4a04b156eeccf30fcfedf4175f09edbc1c407728/test/system/helpers.bash#L8))
         #   `minikube start' failed with status 57
         # * minikube v1.27.0 on Fedora 36 (kvm/amd64)
         # ! Kubernetes 1.25.0 has a known issue with resolv.conf. minikube is using a workaround that should work for most use cases.
         # ! For more information, see: https://github.com/kubernetes/kubernetes/issues/112135
         # * Using the podman driver based on user configuration
         # * The "podman" driver should not be used with root privileges. If you wish to continue as root, use --force.
         # * If you are running minikube within a VM, consider using --driver=none:
         # *   https://minikube.sigs.k8s.io/docs/reference/drivers/none/
         #
         # X Exiting due to DRV_AS_ROOT: The "podman" driver should not be used with root privileges.

Not something I can help with, sorry. Or at least, easily help with. Am posting in case it's something trivial for you to identify and fix. If it isn't, lmk and I'll spin up a VM.

[edsantiago]

The sys failure is a flake, I've filed #16075 for it but there's no point in rerunning it. Anyhow, don't sweat that one.

[umohnani8]

@cevich can minikube be run rootless? root privilege are not required for it when using podman

[umohnani8]

All tests are finally green! @cevich @edsantiago can I please get a final lgtm to merge.

[edsantiago]

git range-diff confirms that the only changes since my last review are rootless and a rebase... so I haven't done a thorough re-review. Feel free to release hold, or wait for another set of eyes.

/lgtm
/hold

[rhatdan]

/hold cancel

[cevich]

Nice work on this @umohnani8 Thanks for seeing it through.