-
Notifications
You must be signed in to change notification settings - Fork 38.6k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
remove [Conformance] flag on some e2es #34140
Conversation
I'm not a big fan. Is anyone against just removing the conformance tag from these tests? |
/cc @kubernetes/sig-testing for comments. |
I'm open to alternatives, but I don't think some of these tests should remove [Conformance] from an upstream perspective. It's just that they're not guaranteed to function downstream given the description above. |
I'm also afraid that we're matching the regex |
We've federated out repos, I'm not entirely certain all the places to look. |
@spxtr This is a priority for us, would [Variant] be ok? |
We don't actually use the conformance label anywhere in the kubernetes repo, so the change doesn't make a big deal to us, but it smells. |
I think there are opinionated differences in deployments that can not easily be resolved. The problem today is upstream only tests/validates 1 configuration, and downstream distributions (e.g. Openshift) have made other choices which have now made the test no longer a [Conformance] test for that environment. So I'm ok with the following:
|
Can you give an example of an opinionated difference? For instance, |
Is it possible to do some sort of |
We lock down our node tightly to not expose any potential attack surface.
No, Moving the change to [Variant], keeps all the current testing conventions with little/no negative upstream impact, what is the logical argument against it? |
The argument is just that |
The 3 options provided earlier...
|
Why is provider not an option? |
Provider in the tests, and other areas, is linked to cloud-provider, which openshift supports many. So provider has an known implementation, and connotation in the code. |
It sounds like we should remove the conformance label, then. This comes up every now and then in @kubernetes/sig-testing. |
subject to change downstream.
a325408
to
eb29c61
Compare
Done. |
I'm fine with this. Anyone from @kubernetes/sig-testing disagree? If not I'll LGTM tomorrow. |
Jenkins GCE e2e failed for commit eb29c61. Full PR test history. The magic incantation to run this job again is |
looks like infra-flakes. |
@k8s-bot test this |
Jenkins GCI GKE smoke e2e failed for commit eb29c61. Full PR test history. The magic incantation to run this job again is |
@k8s-bot test this [submit-queue is verifying that this PR is safe to merge] |
Automatic merge from submit-queue |
Downstream distributions that absorb the upstream tests would like to give their customers a standard mechanism to validate their clusters, post setup. As of today [Conformance] works for most things, but there are a known set of tests that vary due to opinionated differences around networking, security, etc... and providing a complete skip list can be cumbersome. To address this, we've simply modified the flag on some tests to [Conformance:Variant]. All existing behavior should be maintained.
Fixes: #34105
This change is