In the v2 implementation, headless worker service is referenced by jobname-worker-0.jobname-worker (https://github.com/alculquicondor/mpi-operator/blob/master/v2/pkg/controller/mpi_job_controller.go#L1110) which requires the pod /etc/resolv.conf config having namespace.svc.cluster.local presented.
The tf-operator on the other side composes TFCONFIG with both namespace and svc suffix: https://github.com/kubeflow/training-operator/blob/master/pkg/controller.v1/tensorflow/tensorflow.go#L159
Is it possible to consolidate the logic so that the same cluster DNS settings work for both operators?