Skip to content

Install Workload

Install Custom Three-Layer Application

This section deploys a three-layer application (React frontend, Django backend, Postgres database) based on this blog, adapted for Kubernetes and NDB.

Create Database Schema and Data

  1. Apply application secrets: 
    kubectl apply -f https://raw.githubusercontent.com/nutanix-japan/ocp-gitp/main/docs/ocp_ndb/k8s/app-secrets.yaml -n ndb
  2. Download and edit the ConfigMap: 
    curl -LO https://raw.githubusercontent.com/nutanix-japan/ocp-gitp/main/docs/ocp_ndb/k8s/app-variables.yaml
    Update DB_HOST and DB_PORT in app-variables.yaml: 
    DB_HOST: dbforflower-svc
DB_PORT: "80"
  3. Apply the ConfigMap: 
    kubectl apply -f app-variables.yaml -n ndb
  4. Run the Django job to populate the database: 
    kubectl apply -f https://raw.githubusercontent.com/nutanix-japan/ocp-gitp/main/docs/ocp_ndb/k8s/django-job.yaml -n ndb
  5. Monitor the job: 
    kubectl get job django-job -n ndb -w
    Example output: 
    NAME         COMPLETIONS   DURATION   AGE
django-job   1/1           19s        20s
  6. Check job logs: 
    kubectl logs -n ndb $(kubectl get pod -n ndb -l job-name=django-job -o jsonpath='{.items[0].metadata.name}')
  7. Verify new tables: 
    kubectl exec -it psql -n ndb -- psql -h dbforflower-svc -p 80 -U postgres -d predictiondb

    \dt
    Expected output: ~11 tables (e.g., auth_user, django_migrations).

Install Frontend and Backend

  1. Deploy Django and React applications: 
    kubectl apply -f https://raw.githubusercontent.com/nutanix-japan/ocp-gitp/main/docs/ocp_ndb/k8s/django-deployment.yaml -n ndb
kubectl apply -f https://raw.githubusercontent.com/nutanix-japan/ocp-gitp/main/docs/ocp_ndb/k8s/react-deployment.yaml -n ndb
  2. Verify pods: 
    kubectl get pods -n ndb

Create Ingress for Access

  1. Create an Ingress resource to replace OpenShift Routes:

    cat << EOF > ingress.yaml
apiVersion: networking.k8s.io/v1
kind: Ingress
metadata:
  name: flower-app
  namespace: ndb
  annotations:
    traefik.ingress.kubernetes.io/router.entrypoints: web
    traefik.ingress.kubernetes.io/router.pathmatcher: PathPrefix
spec:
  rules:
  - host: flower.apps.k8suserXX.ntnxlab.local
    http:
      paths:
      - path: /
        pathType: Prefix
        backend:
          service:
            name: react-cluster-ip-service
            port:
              number: 80
      - path: /admin
        pathType: Prefix
        backend:
          service:
            name: django-cluster-ip-service
            port:
              number: 80
      - path: /api
        pathType: Prefix
        backend:
          service:
            name: django-cluster-ip-service
            port:
              number: 80
      - path: /static/admin/
        pathType: Prefix
        backend:
          service:
            name: django-cluster-ip-service
            port:
              number: 80
      - path: /static/rest_framework/
        pathType: Prefix
        backend:
          service:
            name: django-cluster-ip-service
            port:
              number: 80
      - path: /static/
        pathType: Prefix
        backend:
          service:
            name: django-cluster-ip-service
            port:
              number: 80
      - path: /media/
        pathType: Prefix
        backend:
          service:
            name: django-cluster-ip-service
            port:
              number: 80
EOF

  2. Replace k8suserXX with your user ID:

    sed -i 's/k8suserXX/your_user_id/g' ingress.yaml

  3. Apply the Ingress:

    kubectl apply -f ingress.yaml -n ndb

  4. Verify if Ingress resources are created

    kubectl get ingress -n ndb

Test Frontend React Application

  1. Update your local hosts file: 
    10.38.18.220 flower.apps.k8suserXX.ntnxlab.local
  2. Access the React app: 
    http://flower.apps.k8suserXX.ntnxlab.local/
  3. Log in with:
    • Username: admin
    • Password: admin_password
  4. Use the sliders to predict flower names.

Test Backend Django Application

  1. Access the Django admin: 
    http://flower.apps.k8suserXX.ntnxlab.local/admin
  2. Log in with:
    • Username: admin
    • Password: admin_password
  3. Add a new user:
    • Username: xyz-user
    • Password: your_password
  4. Save and test the new user in the React app.

Check Postgres Database Data

  1. Connect to the database: 
    kubectl exec -it psql -n ndb -- psql -h dbforflower-svc -p 80 -U postgres -d predictiondb
  2. Query users: 
    SELECT username, last_login FROM auth_user;
    Expected output: 
    username  | last_login
----------+-------------------------------
admin     | 2022-12-14 01:38:41.480801+00
xyz-user  | 2022-12-14 01:38:53.474404+00

Takeaways

  • The NDB Operator simplifies VM-based database deployment in Kubernetes.
  • Databases are provisioned via YAML manifests and exposed as Kubernetes Services.
  • Security is managed with securityContext instead of SCCs, ensuring non-root execution.

Cleanup (optional)

kubectl delete -f ingress.yaml -n ndb
kubectl delete -f https://raw.githubusercontent.com/nutanix-japan/ocp-gitp/main/docs/ocp_ndb/k8s/react-deployment.yaml -n ndb
kubectl delete -f https://raw.githubusercontent.com/nutanix-japan/ocp-gitp/main/docs/ocp_ndb/k8s/django-deployment.yaml -n ndb
kubectl delete -f https://raw.githubusercontent.com/nutanix-japan/ocp-gitp/main/docs/ocp_ndb/k8s/django-job.yaml -n ndb
kubectl delete -f app-variables.yaml -n ndb
kubectl delete -f https://raw.githubusercontent.com/nutanix-japan/ocp-gitp/main/docs/ocp_ndb/k8s/app-secrets.yaml -n ndb
kubectl delete -f database.yaml -n ndb
kubectl delete -f ndbserver.yaml -n ndb
kubectl delete -f your-db-secret.yaml -n ndb
kubectl delete -f your-ndb-secret.yaml -n ndb
kubectl delete pod psql -n ndb
kubectl delete namespace ndb
helm uninstall ndb-operator -n ndb-operator
kubectl delete namespace ndb-operator

Decommission the Postgres database in the NDB UI.