Step 3: Install Workflow Components

Digital Fingerprinting

This section will walk through an end-to-end workflow deployment using the example software stack components previously described.

Ensure that the previous Step 2: Set Up Required Infrastructure section has been completed prior to proceeding with the deployment steps.

1. SSH into the cluster.

2. If you still need to install and set up the NGC CLI with your API Key and NGC Org information, please do so to download the resources necessary for the workflow. Instructions can be found here.

3. Pull the helm chart onto the system using the following command:

   Copy

   Copied!

               
   
               
   helm fetch https://helm.ngc.nvidia.com/nvaie/charts/cybersecurity-dfp-0.2.1.tgz --username='$oauthtoken' --password=<YOUR API KEY>
           

   
   

4. Next, set the following environment variables according to your information as shown in the example below:

   Copy

   Copied!

               
   
               
   export NGC_KEY=<YOUR API KEY>
   export NAMESPACE=<A namespace name, for example, “aiworkflow”>
   export APP_NAME=<An app name to use within the namespace, for example, “dfp”>
           

   
   

   The content of these environment variables will be referenced throughout the rest of the workflow.

5. Once the environment variables are set, run the Helm install command below to deploy the application.

   Copy

   Copied!

               
   
               
   helm install -n $NAMESPACE $APP_NAME ./cybersecurity-dfp-0.2.1.tgz --set ngcKey=$NGC_KEY --create-namespace --timeout 3600s
           

   Note

   This installation may take a few minutes to complete.

   Note

   This Helm chart assumes that the default configuration described in the previous steps has been installed, including NVIDIA Cloud Native Service Add-on Pack and the local-path-provisioner. If the cluster configuration differs, such as if a different storage class is created, the values.yaml for the chart will need to be modified accordingly.

   
   

6. After a few minutes, check and verify that the application deployed successfully using the following command:

   Copy

   Copied!

               
   
               
   kubectl get pods -n $NAMESPACE
           

   Note

   You may see a mlflow-backup pod with a status of Error. If another mlflow-backup pod has a status of Completed, this is benign and can be ignored.

   
   

   

7. At this point, proceed to the Running the Workflow section to continue.