mpijob-bibw.yaml

apiVersion: kubeflow.org/v2beta1
kind: MPIJob
metadata:
  label:
    profile: rootless
  name: osu-mpi-bibw
spec:
  sshAuthMountPath: /home/mpi/.ssh
  cleanPodPolicy: Running
  slotsPerWorker: 1
  mpiReplicaSpecs:
    Launcher:
      replicas: 1
      template:
        spec:
          initContainers:
          - name: wait-hostfilename
            image: image-registry.openshift-image-registry.svc:5000/mpi-benchmark/mpi:base
            command:
            - bash
            - -cx
            - "[[ $(cat /etc/mpi/hostfile | wc -l) != 0 ]] && (date; echo 'Hostfile is ready'; cat /etc/mpi/hostfile) || (date; echo 'Hostfile not ready ...'; sleep 10; exit 1) && while read host; do while ! ssh $host echo $host ; do date; echo \"Pod $host is not up ...\"; sleep 10; done; date; echo \"Pod $host is ready\"; done <<< \"$(cat /etc/mpi/hostfile)\""
            volumeMounts:
            - mountPath: /etc/mpi
              name: mpi-job-config
            - mountPath: /home/mpi/.ssh
              name: ssh-auth
          containers:
          - name: mpi-launcher
            command:
            - mpirun
            - --allow-run-as-root
            - -np
            - "2"
            - -bind-to
            - none
            - -map-by
            - slot
            - -mca
            - pml
            - ob1
            - -mca
            - btl
            - ^openib
            - /opt/osu-micro-benchmarks/libexec/osu-micro-benchmarks/mpi/pt2pt/osu_bibw
            image: image-registry.openshift-image-registry.svc:5000/mpi-benchmark/mpi:base
            imagePullPolicy: Always
    Worker:
      replicas: 2
      template:
        metadata:
          labels:
            app: mpi-app
        spec:
          topologySpreadConstraints:
          - maxSkew: 1
            topologyKey: kubernetes.io/hostname
            whenUnsatisfiable: DoNotSchedule
            labelSelector:
              matchLabels:
                app: mpi-app
          containers:
          - name: mpi-worker
            image: image-registry.openshift-image-registry.svc:5000/mpi-benchmark/mpi:osu-mpi
            imagePullPolicy: Always