01_Ray_setup.md

Ray Setup

TOC

• 1. Install KubeRay operator
• 2. Install Ray cluster
• 3. Connect to Ray and run a job (via Kubeflow)
  • 3.1 Forward the port of Istio’s Ingress-Gateway
  • 3.2 Create a JupyterLab via Kubeflow Central Dashboard
  • 3.3 Run a Ray job
• 4. Access Ray dashboard
• 5. Connect to Ray cluster from local machine

Prerequisites

• Helm

  # you can use the install_helm.sh script to install Helm if you haven't already
  bash scripts/install_helm.sh

1. Install KubeRay operator

helm repo add kuberay https://ray-project.github.io/kuberay-helm/
helm repo update

# Install both CRDs and KubeRay operator v1.0.0.
helm install kuberay-operator kuberay/kuberay-operator --version 1.0.0

# Confirm that the operator is running in the namespace `default`.
kubectl get pods
# NAME                                READY   STATUS    RESTARTS   AGE
# kuberay-operator-7fbdbf8c89-pt8bk   1/1     Running   0          27s

2. Install Ray cluster

# Create a RayCluster CR, and the KubeRay operator will reconcile a Ray cluster
# with 1 head Pod and 1 worker Pod.
helm install raycluster kuberay/ray-cluster --version 1.0.0 --set image.tag=2.7.0

# Check RayCluster
kubectl get pod -l ray.io/cluster=raycluster-kuberay
# NAME                                          READY   STATUS    RESTARTS   AGE
# raycluster-kuberay-head-bz77b                 1/1     Running   0          64s
# raycluster-kuberay-worker-workergroup-8gr5q   1/1     Running   0          63s

To customize the Ray cluster (e.g. number of worker Pods), you can modify the default Helm values:

helm show values kuberay/ray-cluster > values.yaml
# Modify the values.yaml, then install the Ray cluster with the customized values.
helm install raycluster kuberay/ray-cluster -f values.yaml --version 1.0.0 --set image.tag=2.7.0

For more details, please refer to Ray-Kubeflow documentation.