ray-cluster.tpu-v4-singlehost.yaml

# This template contains a Kuberay cluster using a 2x2x1 TPU v4 PodSlice.
# To get access to TPU resources, please follow instructions in this link:
# https://cloud.google.com/kubernetes-engine/docs/how-to/tpus
apiVersion: ray.io/v1
kind: RayCluster
metadata:
  # Label required for TPU webhook to initialize environments.
  labels:
    app.kubernetes.io/name: kuberay
  name: example-cluster-kuberay
spec:
  headGroupSpec:
    rayStartParams:
      {}
    template:
      spec:
        imagePullSecrets:
          []
        containers:
          - volumeMounts:
            - mountPath: /tmp/ray
              name: ray-logs
            name: ray-head
            image: rayproject/ray:2.9.0-py310
            imagePullPolicy: IfNotPresent
            resources:
              limits:
                cpu: "8"
                ephemeral-storage: 20Gi
                memory: 40G
              requests:
                cpu: "8"
                ephemeral-storage: 10Gi
                memory: 40G
            securityContext:
              {}
            env:
              - name: RAY_memory_monitor_refresh_ms
                value: "0"
              - name: RAY_GRAFANA_IFRAME_HOST
                value: http://${grafana_host}
              - name: RAY_GRAFANA_HOST
                value: http://grafana:80
              - name: RAY_PROMETHEUS_HOST
                value: http://frontend:9090
            ports:
              - containerPort: 6379
                name: gcs
              - containerPort: 8265
                name: dashboard
              - containerPort: 10001
                name: client
              - containerPort: 8000
                name: serve
        volumes:
          - emptyDir: {}
            name: ray-logs
      metadata:
        labels:
          cloud.google.com/gke-ray-node-type: head
          app.kubernetes.io/name: kuberay
          app.kubernetes.io/instance: example-cluster

  workerGroupSpecs:
  - rayStartParams:
      {}
    replicas: 1
    minReplicas: 1
    maxReplicas: 1
    numOfHosts: 1
    groupName: workergroup
    template:
      spec:
        imagePullSecrets:
          []
        containers:
          - volumeMounts:
            - mountPath: /tmp/ray
              name: ray-logs
            name: ray-worker
            image: rayproject/ray:2.9.0-py310
            imagePullPolicy: IfNotPresent
            resources:
              limits:
                cpu: "1"
                ephemeral-storage: 20Gi
                google.com/tpu: "4"
                memory: 40G
              requests:
                cpu: "1"
                ephemeral-storage: 10Gi
                google.com/tpu: "4"
                memory: 40G
            securityContext:
              {}
            env:
            ports:
              null
        volumes:
          - emptyDir: {}
            name: ray-logs
        nodeSelector:
          cloud.google.com/gke-tpu-accelerator: tpu-v4-podslice
          cloud.google.com/gke-tpu-topology: 2x2x1
      metadata:
        labels:
          cloud.google.com/gke-ray-node-type: worker
          app.kubernetes.io/name: kuberay
          app.kubernetes.io/instance: example-cluster