Fix crash in podgroup when runLauncherAsWorker is true

[GonzaloSaez]

When runLauncherAsWorker is true and there is no worker, the MPI controller will crash in calPGMinResource

[google-oss-prow]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please assign rongou for approval. For more information see the Kubernetes Code Review Process.

The full list of commands accepted by this bot can be found here.

Needs approval from an approver in each of these files:

• OWNERS

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

[kuizhiqing]

lgtm

[tenzen-y]

Thank you for fixing this! Could you add a unit test case in

mpi-operator/pkg/controller/podgroup_test.go

Line 369 in c738a83

func TestCalculatePGMinResources(t *testing.T) {

?

[tenzen-y]

@GonzaloSaez Could you sign DCO (https://github.com/kubeflow/mpi-operator/pull/669/checks?check_run_id=32329697546), and address CI errors?

[GonzaloSaez]

Sure will do. Are these tests flaky? It seems that only IntelMPI is failing and I don't see why my changes would make those tests fail

[tenzen-y]

Sure will do. Are these tests flaky? It seems that only IntelMPI is failing and I don't see why my changes would make those tests fail

Uhm interesting. Let me restart Jobs.

[tenzen-y]

Sure will do. Are these tests flaky? It seems that only IntelMPI is failing and I don't see why my changes would make those tests fail

@GonzaloSaez I restarted the CI three times, all trials failed. So, this PR seems to bring any kind of additional bug. Could you fix that?

[GonzaloSaez]

@tenzen-y I ran these tests locally and they are passing

------------------------------
Deleting cluster "kind" ...
Deleted nodes: ["kind-control-plane"]

Ran 13 of 13 Specs in 725.625 seconds
SUCCESS! -- 13 Passed | 0 Failed | 0 Pending | 0 Skipped
--- PASS: TestE2E (725.63s)
PASS
ok      github.com/kubeflow/mpi-operator/test/e2e       725.655s

I'm still looking in to why this is happening in CI.

[tenzen-y]

@tenzen-y I ran these tests locally and they are passing

------------------------------
Deleting cluster "kind" ...
Deleted nodes: ["kind-control-plane"]

Ran 13 of 13 Specs in 725.625 seconds
SUCCESS! -- 13 Passed | 0 Failed | 0 Pending | 0 Skipped
--- PASS: TestE2E (725.63s)
PASS
ok      github.com/kubeflow/mpi-operator/test/e2e       725.655s

I'm still looking in to why this is happening in CI.

Interesting. Let's see if the master branch is health here: #671

[tenzen-y]

@tenzen-y I ran these tests locally and they are passing

------------------------------
Deleting cluster "kind" ...
Deleted nodes: ["kind-control-plane"]

Ran 13 of 13 Specs in 725.625 seconds
SUCCESS! -- 13 Passed | 0 Failed | 0 Pending | 0 Skipped
--- PASS: TestE2E (725.63s)
PASS
ok      github.com/kubeflow/mpi-operator/test/e2e       725.655s

I'm still looking in to why this is happening in CI.

Interesting. Let's see if the master branch is health here: #671

It seems the master branch E2E has been broken ...