Deployment Guide¶

This guide covers deploying NITA in production, managing pod lifecycles, upgrading components, and operational best practices.

Deployment Model¶

NITA deploys as a single-node Kubernetes cluster with all components running within the nita namespace.

graph TB
    subgraph "Production Host"
        subgraph "Kubernetes (kubeadm)"
            subgraph "nita namespace"
                P["🛡️ Proxy :443"]
                W["🌐 Webapp :8000"]
                J["⚙️ Jenkins :8443"]
                D["🗄️ DB :3306"]
            end
            CALICO["Calico CNI"]
        end
        CONTAINERD["containerd runtime"]
        OS["Ubuntu 22.04 / AlmaLinux 9.3"]
    end

    USERS["👥 Users"] -->|"HTTPS"| P
    P --> W
    P --> J
    W --> D

Deployment Procedure¶

Fresh Deployment¶

Run the automated Installation script:

sudo -E ./install.sh

After installation, verify all pods are running:

nita-cmd kube pods

Access Points¶

Upgrading Components¶

Upgrade a Container Image¶

1. Build or pull the new image

2. Import into containerd:

   docker save juniper/nita-jenkins:<new-tag> > nita-jenkins.tar
   sudo ctr -n=k8s.io image import nita-jenkins.tar
   

3. Update the deployment YAML:

   sudo vi /opt/nita/k8s/jenkins-deployment.yaml
   # Change the image: tag
   

4. Apply the change:

   kubectl delete deployment jenkins -n nita
   kubectl apply -f /opt/nita/k8s/jenkins-deployment.yaml
   

Upgrade Kubernetes¶

NITA uses Kubernetes 1.28. To upgrade:

# Update kubeadm
sudo apt install -y kubeadm=<new-version>

# Plan the upgrade
sudo kubeadm upgrade plan

# Apply the upgrade
sudo kubeadm upgrade apply v<new-version>

# Update kubelet and kubectl
sudo apt install -y kubelet=<new-version> kubectl=<new-version>
sudo systemctl restart kubelet

Operational Procedures¶

Start / Stop NITA¶

Stop all NITA pods:

kubectl scale deployment --all --replicas=0 -n nita

Start all NITA pods:

kubectl scale deployment --all --replicas=1 -n nita

Restart a Single Component¶

kubectl rollout restart deployment/<name> -n nita

Health Checks¶

# Pod status
kubectl get pods -n nita

# Pod events / errors
kubectl describe pod <pod-name> -n nita

# Container logs
kubectl logs <pod-name> -n nita

# Resource usage
nita-cmd stats

Backup & Restore¶

Jenkins Backup¶

Jenkins data is stored in the jenkins-home PVC (20 Gi):

# Backup Jenkins jobs and configuration
kubectl exec -it -n nita <jenkins-pod> -- tar czf /tmp/jenkins-backup.tar.gz /var/jenkins_home
kubectl cp nita/<jenkins-pod>:/tmp/jenkins-backup.tar.gz ./jenkins-backup.tar.gz

Or use nita-cmd:

nita-cmd jenkins backup
nita-cmd jenkins restore

MariaDB Backup¶

# Dump the database
kubectl exec -it -n nita <db-pod> -- mysqldump -u root -proot Sites > sites-backup.sql

# Restore
kubectl exec -i -n nita <db-pod> -- mysql -u root -proot Sites < sites-backup.sql

Security Considerations¶

Monitoring¶

Pod Resource Usage¶

nita-cmd stats

Example output:

CONTAINER ID   NAME                    CPU %   MEM USAGE / LIMIT   MEM %   NET I/O         BLOCK I/O
34743b207365   nitawebapp_webapp_1     5.85%   88.74MiB / 3.853GiB 2.25%   50.8kB / 25.8kB 44.9MB / 0B
922a4eb21e05   nitajenkins_jenkins_1   0.08%   670.4MiB / 3.853GiB 16.99%  3.09MB / 52.8kB 154MB / 6.94MB

Kubernetes Dashboard (Optional)¶

You can deploy the Kubernetes dashboard for a web-based view:

kubectl apply -f https://raw.githubusercontent.com/kubernetes/dashboard/v2.7.0/aio/deploy/recommended.yaml

Network Requirements¶

Architecture Decision: Single-Node¶

NITA is designed for single-node Kubernetes deployments. This simplifies:

• Installation and management
• Persistent volume mounting (hostPath)
• Network complexity (all internal)

For multi-node deployments, you would need:

• Shared storage (NFS, Ceph, etc.)
• Network-aware PVC provisioners
• Load balancer for proxy