hack/ginkgo-e2e.sh fails after upgrading to ginkgo v2

[nilo19]

Which jobs are failing?

After #109111 was merged, hack/ginkgo-e2e.sh fails because of:

I0708 07:50:08.498104 93818 e2e.go:116] Starting e2e run "99821dc1-5e1d-48c9-8be4-303d1ba74adb" on Ginkgo node 1
Ginkgo detected configuration issues:
Use of go test -parallel
Go test's implementation of parallelization does not actually parallelize
Ginkgo specs. Please use the ginkgo cli and ginkgo -p or ginkgo -procs=N
instead.

Did I miss some configs or it is a regression?

Which tests are failing?

hack/ginkgo-e2e.sh

Since when has it been failing?

After #109111 was merged.

Testgrid link

https://prow.k8s.io/view/gs/kubernetes-jenkins/pr-logs/pull/kubernetes-sigs_cloud-provider-azure/1991/pull-cloud-provider-azure-e2e-capz/1545557471236460544

Reason for failure (if possible)

I0708 07:50:08.498104 93818 e2e.go:116] Starting e2e run "99821dc1-5e1d-48c9-8be4-303d1ba74adb" on Ginkgo node 1
Ginkgo detected configuration issues:
Use of go test -parallel
Go test's implementation of parallelization does not actually parallelize
Ginkgo specs. Please use the ginkgo cli and ginkgo -p or ginkgo -procs=N
instead.

Anything else we need to know?

cc @chendave

Relevant SIG(s)

/sig testing

[nilo19]

/triage accpted

[k8s-ci-robot]

@nilo19: The label(s) triage/accpted cannot be applied, because the repository doesn't have them.

In response to this:

/triage accpted

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.

[nilo19]

/triage accepted

[aojea]

the error mentions go test but the hack/ginkgo-e2e.sh script seems to run ginkgo directly 🤔

[nilo19]

In our code we just call ginkgo-e2e.sh directly. https://github.com/kubernetes-sigs/cloud-provider-azure/blob/master/hack/test_k8s_e2e_capz.sh#L52

[chendave]

Use of go test -parallel

This was checked in Ginkgo, https://github.com/onsi/ginkgo/blob/master/types/errors.go#L524-L525

I am not familiar with the test framework of that project (cloud-provider-azure), I did a quick search, and found the script hack/run-e2e-tests.sh called go test directly, pls see code, can you pls check how this script got called, and is there a parallel flag passed to that script?

[chendave]

okay, I managed to reproduce the problem,

/usr/bin/e2e.test -ginkgo.focus="\[sig-network\] ClusterDns \[Feature:Example\] should create pod that uses dns" -test.parallel=30 -report-dir=/var/log/kubernetes/e2e
Jul 11 11:14:32.809: INFO: The --provider flag is not set. Continuing as if --provider=skeleton had been used.
I0711 11:14:32.809811    5754 e2e.go:118] Starting e2e run "0cba6b71-ad52-4b38-a979-6a873a8964de" on Ginkgo node 1
Ginkgo detected configuration issues:
Use of go test -parallel
  Go test's implementation of parallelization does not actually parallelize
  Ginkgo specs.  Please use the ginkgo cli and `ginkgo -p` or `ginkgo -procs=N`
  instead.

pls don't use that parameter, --test.parallel=30

https://github.com/kubernetes/test-infra/blob/master/config/jobs/kubernetes-sigs/cloud-provider-azure/cloud-provider-azure-config.yaml#L182

And switch to the parameter provided by Ginkgo.

[chendave]

And as it said, you can make it by ginkgo -p or ginkgo -procs=N, this should be a restriction introduced by Ginkgo v2.

@pohly I think we should add this to the action list of #109111.