ray-k8s部署

u0119608232026-05-27 22:23

https://docs.ray.io/en/latest/cluster/kubernetes/getting-started/raycluster-quick-start.html#kuberay-raycluster-quickstart

Deploy a KubeRay operator

bash 复制代码

helm repo add kuberay https://ray-project.github.io/kuberay-helm/
helm repo update
# Install both CRDs and KubeRay operator v1.6.0.
helm install kuberay-operator kuberay/kuberay-operator --version 1.6.0

bash 复制代码

kubectl get pods

Deploy a RayCluster custom resource

bash 复制代码

image:
  repository: rayproject/ray
  tag: "2.55.1-py311-aarch64"
  pullPolicy: IfNotPresent
head:
  rayStartParams:
    dashboard-host: "0.0.0.0"
    object-store-memory: "200000000"
  resources:
    requests:
      cpu: "500m"
      memory: "1536Mi"
    limits:
      cpu: "2"
      memory: "3072Mi"
worker:
  replicas: 1
  minReplicas: 0
  maxReplicas: 1
  rayStartParams:
    object-store-memory: "200000000"
  resources:
    requests:
      cpu: "500m"
      memory: "1024Mi"
    limits:
      cpu: "2"
      memory: "2048Mi"

bash 复制代码

helm install raycluster kuberay/ray-cluster --version 1.6.0
helm install raycluster kuberay/ray-cluster --version 1.6.0 -f raycluster-k3d-small-values.yaml

# Once the RayCluster CR has been created, you can view it by running:
kubectl get rayclusters

# View the pods in the RayCluster named "raycluster-kuberay"
kubectl get pods --selector=ray.io/cluster=raycluster-kuberay

端口转发

bash 复制代码

kubectl port-forward -n prefect svc/raycluster-kuberay-head-svc 8265:8265 10001:10001 --address 0.0.0.0