Deploying NVIDIA AI Enterprise Containers

1. Create a Image Pull secret with your NGC key

   oc create secret docker-registry regcred --docker-server=nvcr.io/nvaie --docker-username=\$oauthtoken --docker-password=<YOUR_NGC_KEY> --docker-email=<your_email_id> -n default
   

2. Apply the custom yaml files accessing NVAIE Jupyter notebooks

   In this example a TensorFlow container is leveraged to launch a Jupyter notebook. Create a new yaml file with contents below called tensorflow.yaml

   Warning

   Where <NVAIE-MAJOR-VERSION>:<NVAIE-CONTAINER-TAG> fill the version and tag. Example: 2-3:22.09-tf2-nvaie-2.3-py3

   apiVersion: apps/v1
   kind: Deployment
   metadata:
   name: tensorflow-jupyter-notebook
   labels:
       app: tensorflow-jupyter-notebook
   spec:
   replicas: 1
   selector:
       matchLabels:
       app: tensorflow-jupyter-notebook
   template:
       metadata:
       labels:
           app: tensorflow-jupyter-notebook
       spec:
       containers:
       - name: tensorflow-container
           image: nvcr.io/nvaie/tensorflow-<NVAIE-MAJOR-VERSION>:<NVAIE-CONTAINER-TAG>
           ports:
           - containerPort: 8888
           command: ["jupyter-notebook"]
           args: ["--NotebookApp.token=''"]
           resources:
           requests:
               nvidia.com/gpu: 1
           limits:
               nvidia.com/gpu: 1
       imagePullSecrets:
       - name: regcred
   ---
   kind: Service
   apiVersion: v1
   metadata:
   name: tensorflow-jupyter-notebook
   spec:
   type: NodePort
   selector:
       app: tensorflow-jupyter-notebook
   ports:
   - protocol: TCP
       nodePort: 30040
       port: 8888
       targetPort: 8888
   

   Warning

   Leverage routes in instead of Node Port for production deployments. Leverage the Red Hat Route Configuration document for route creation.

   Run the command below to apply the yaml file.

   oc apply -f tensorflow.yaml
   

3. Access the Jupyter notebook

   Once the pod is started describe it with the following:

   oc describe pod tensorflow-jupyter-notebook-6c87cb86f7-9hvpm
   

   Note the FQDN or IP of the node it is running on and construct the URL for accessing the notebook.

   http://<NODE_FQDN_OR_IP>:30040
   

   Note

   Additional information for launching Jupyter Notebooks with OpenShift can be found in the Using Jupyter Project Images blog. Additional examples can be found in the Jupyter-Stack OpenShift Exampels Gitlab page.