Porting v1alpha2 job to v2beta1 on Google Kubernetes Engine

Hi guys,

I am currently working on a research project about using Kubernetes to run mpi jobs on containers as hosts for teaching parallel and distributed computing. Our main platform right now is Google Kubernetes Engine. Previously, I used version V1alpha2 and it worked flawlessly. However, since I changed to version V2beta1 now, the mpi-operator is acting up. The problem is when I use command:

kubectl get pods

Neither the launcher pod nor worker pods showed up with V2beta1, as they did in V1. The configuration for the pods is as follow:

Are there examples or guidance you can point me to, in order to port my working V1 configuration to V2?

Start with the V1alpha2 and V2beta1 configurations shown in the issue and reproduce the behavior with kubectl get pods on Google Kubernetes Engine. Compare the available mpi-operator examples or migration guidance for these versions, then confirm that both launcher and worker pods are created for the V2beta1 configuration.

Written by the indexing model from the issue text.