Submarine's Helm Chart will deploy Submarine Server, TF/PyTorch Operator, Notebook controller and Traefik. We use the TF/PyTorch operator to run tf/pytorch job, the notebook controller to manage jupyter notebook and Traefik as reverse-proxy.
Helm v3 is minimum requirement. See here for installation: https://helm.sh/docs/intro/install/
The Submarine helm charts is released with the source code for now. Please go to http://submarine.apache.org/download.html to download
cd <PathTo>/submarine helm install submarine ./helm-charts/submarine
This will install submarine in the “default” namespace. The images are from Docker hub apache/submarine. See ./helm-charts/submarine/values.yaml for more details
If we'd like use a different namespace like “submarine”
kubectl create namespace submarine helm install submarine ./helm-charts/submarine -n submarine
Note that if you encounter below issue when installation:
Error: rendered manifests contain a resource that already exists. Unable to continue with install: existing resource conflict: namespace: , name: podgroups.scheduling.incubator.k8s.io, existing_kind: apiextensions.k8s.io/v1beta1, Kind=CustomResourceDefinition, new_kind: apiextensions.k8s.io/v1beta1, Kind=CustomResourceDefinition
It might be caused by the previous installed submarine charts. Fix it by running:
kubectl delete crd/tfjobs.kubeflow.org && kubectl delete crd/podgroups.scheduling.incubator.k8s.io && kubectl delete crd/pytorchjobs.kubeflow.org
Once you got it installed, check with below commands and you should see similar outputs:
kubectl get pods
NAME READY STATUS RESTARTS AGE notebook-controller-deployment-5db8b6cbf7-k65jm 1/1 Running 0 5s pytorch-operator-7ff5d96d59-gx7f5 1/1 Running 0 5s submarine-database-8d95d74f7-ntvqp 1/1 Running 0 5s submarine-server-b6cd4787b-7bvr7 1/1 Running 0 5s submarine-traefik-9bb6f8577-66sx6 1/1 Running 0 5s tf-job-operator-7844656dd-lfgmd 1/1 Running 0 5s
Submarine server by default expose 8080 port within K8s cluster. After Submarine v0.5 uses Traefik as reverse-proxy by default. If you don't want to use Traefik, you can modify below value to false in ./helm-charts/submarine/values.yaml.
# Use Traefik by default traefik: enabled: true
To access the server from outside of the cluster, we use Traefik ingress controller and NodePort for external access.
Please refer to ./helm-charts/submarine/charts/traefik/values.yaml and Traefik docs for more details if you want to customize the default value for Traefik.
# Use nodePort and Traefik ingress controller by default. # To access the submarine server, open the following URL in your browser. http://127.0.0.1:32080
Or you can use port-forward to forward a local port to a port on the submarine server pod.
# Use port-forward kubectl port-forward svc/submarine-server 8080:8080 # In another terminal. Run below command to verify it works curl http://127.0.0.1:8080/api/v1/experiment/ping {"status":"OK","code":200,"success":true,"message":null,"result":"Pong","attributes":{}}
helm delete submarine