KubeRay CLI
KubeRay CLI provides the ability to manage kuberay resources (ray clusters, compute templates etc) through command line interface.
Installation
Please check release page and download the binaries.
Prerequisites
- Kuberay operator needs to be running.
- Kuberay apiserver needs to be running and accessible.
Development
- Kuberay CLI uses Cobra framework for the CLI application.
- Kuberay CLI depends on kuberay apiserver to manage these resources by sending grpc requests to the kuberay apiserver.
You can build kuberay binary following this way.
cd kuberay/cli
go build -o kuberay -a main.go
Usage
Configure kuberay apiserver endpoint
- Default kuberay apiserver endpoint:
127.0.0.1:8887. - If kuberay apiserver is not run locally, this must be set in order to manage ray clusters and ray compute templates.
Read current kuberay apiserver endpoint
./kuberay config get endpoint
Reset kuberay apiserver endpoint to default (127.0.0.1:8887)
./kuberay config reset endpoint
Set kuberay apiserver endpoint
./kuberay config set endpoint <kuberay apiserver endpoint>
Manage Ray Clusters
Create a Ray Cluster
Usage:
kuberay cluster create [flags]
Flags:
--environment string environment of the cluster (valid values: DEV, TESTING, STAGING, PRODUCTION) (default "DEV")
--head-compute-tempalte string compuate template name for ray head
--head-image string ray head image
--head-service-type string ray head service type (ClusterIP, NodePort, LoadBalancer) (default "ClusterIP")
--name string name of the cluster
--namespace string kubernetes namespace where the cluster will be (default "ray-system")
--user string SSO username of ray cluster creator
--version string version of the ray cluster (default "1.9.0")
--worker-compute-template string compute template name of worker in the first worker group
--worker-group-name string first worker group name
--worker-image string image of worker in the first worker group
--worker-replicas uint32 pod replicas of workers in the first worker group (default 1)
Known Limitation: Currently only one worker compute template is supported during creation.
Get a Ray Cluster
./kuberay cluster get <cluster name>
List Ray Clusters
./kuberay cluster list
Delete a Ray Cluster
./kuberay cluster delete <cluster name>
Manage Ray Compute Template
Create a Compute Template
Usage:
kuberay template compute create [flags]
Flags:
--cpu uint32 ray pod CPU (default 1)
--gpu uint32 ray head GPU
--gpu-accelerator string GPU Accelerator type
--memory uint32 ray pod memory in GB (default 1)
--name string name of the compute template
Get a Ray Compute Template
./kuberay template compute get <compute template name>
List Ray Compute Templates
./kuberay template compute list
Delete a Ray Compute Template
./kuberay template compute delete <compute template name>
End to end example
Configure the endpoints
kubectl port-forward svc/kuberay-apiserver-service 8887:8887 -n ray-system
./kuberay config set endpoint 127.0.0.1:8887
Create compute templates
./kuberay template compute create --cpu 2 --memory 4 --name "worker-template"
./kuberay template compute create --cpu 1 --memory 2 --name "head-template"
List compute templates created
./kuberay template compute list
Create the cluster
./kuberay cluster create --name test-cluster --user jiaxin.shan \
--head-compute-template head-template \
--head-image rayproject/ray:1.9.2 \
--worker-group-name small-wg \
--worker-compute-template worker-template \
--worker-image rayproject/ray:1.9.2
List the clusters
./kuberay cluster list