This change is 

fixes #38569

Jenkins unit/integration failed for commit e4302a2. Full PR test history.

The magic incantation to run this job again is @k8s-bot unit test this. Please help us cut down flakes by linking to an open flake issue when you hit one in your PR.

I don't really understand the logic of func cleanupGCE(gceController *GCEIngressController).

In its first poll loop it calls gceController.Cleanup(false), which will delete nothing and always return err IMO? Because when del is set to false, gcloudDelete() will not be called, and the returned msg will be non-empty?

Yeah in the poll loops it's waiting for the ingress controller to cleanup, and simply logging the remaining resources. After the provided timeout, if the ingress controller still hasn't cleaned up, we go ahead and fail the test, and delete all leaked resources in the test with Cleanup(true).

This is an issue that plagues our e2es. Some controller in the cluster has a bug that leaks some resource. We fail the test, but never actually cleanup the resource, till the gce project is full of quota and the submit queue closes till some human goes in and deletes them. The cleanup(false)/cleanup(true) calls are addressing that.

Make sense. Thanks for explaining.

/lgtm

@k8s-bot unit test this, issue: #33461

@k8s-bot test this [submit-queue is verifying that this PR is safe to merge]

Jenkins GCI GKE smoke e2e failed for commit e4302a2. Full PR test history.

The magic incantation to run this job again is @k8s-bot gci gke e2e test this. Please help us cut down flakes by linking to an open flake issue when you hit one in your PR.

Automatic merge from submit-queue