Skip to main content
Create a Managed Kubernetes cluster for Data Science
Last update:

Create a Managed Kubernetes cluster for Data Science

In a Managed Kubernetes cluster, you can run a container with pre-installed machine learning tools and run the Jupyter Notebook service in it.

Container can be used to train and infer models for application development and data manipulation.

List of tools⁠

A list of packages in the container:

Create a cluster for Data Science⁠

  1. In Control Panel, go to Cloud Platform → Kubernetes.
  2. Click Create Cluster.
  3. Select a node group configuration with parameters of at least 4 vCPUs, 8 GB RAM, 20 GB SSD.
  4. Select the rest of the cluster settings (more details in the Create Managed Kubernetes cluster instructions) and click Create.
  5. Connect to cluster.

Start the container⁠

  1. Download YAML file with deployment configuration.

  2. Start the container:

    kubectl apply -f selectel-ml.yaml

  3. Check the status of the container:

    kubectl get pod -w

  4. Wait for the Running status — it means that the container is created and running:

    selectel-ml 1/1 Running

Launch Jupyter Notebook⁠

  1. Open a port to access the service:

    kubectl expose deployment selectel-ml --type=LoadBalancer --name=my-service

  2. Get the port to connect to the Jupyter server:

    kubectl get services

    Example answer:

    NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE
    my-service LoadBalancer 10.100.90.86 203.0.113.1 8888:31779/TCP 30s

  3. In the address bar of your browser, enter the address from EXTERNAL-IP and the port number from PORT(S), for example 203.0.113.1:8888.

  4. If <pending> is displayed in EXTERNAL-IP, run the kubectl get services command after a few minutes.

  5. In the Jupyter Notebook web interface that opens, enter the default password: 9lG0eXCevt.

  6. Optional: Change the password according to Jupyter Notebook instructions.

Working with the container through the console⁠

Configure the container operation through the console:

kubectl exec -it [pod name] /bin/bash