Skip to content

MINIKUBE_RESTART.md

Latest commit

 

History

History
80 lines (59 loc) · 1.68 KB

MINIKUBE_RESTART.md

File metadata and controls

80 lines (59 loc) · 1.68 KB
Raw

Restarting RAG API in Minikube

Quick Restart Commands

Method 1: Restart Deployment (Recommended)

kubectl rollout restart deployment rag-app-deployment

Method 2: Delete Pod (Auto-recreates)

# Get pod name
POD_NAME=$(kubectl get pods -l app=rag-app -o jsonpath='{.items[0].metadata.name}')

# Delete pod (Kubernetes will automatically recreate it)
kubectl delete pod $POD_NAME

Method 3: Scale Down/Up

# Scale down to 0
kubectl scale deployment rag-app-deployment --replicas=0

# Scale back up to 1
kubectl scale deployment rag-app-deployment --replicas=1

Check Status

# Check deployment status
kubectl get deployment rag-app-deployment

# Check pod status
kubectl get pods -l app=rag-app

# Check service
kubectl get service rag-api-service

# View logs
kubectl logs -l app=rag-app --tail=50 -f

Access the API

After restart, access via:

# Start minikube tunnel (if not already running)
minikube service rag-api-service

# Or access directly via NodePort
curl http://192.168.49.2:30001/

Current Setup

  • Deployment: rag-app-deployment
  • Service: rag-api-service (NodePort, port 30001)
  • Namespace: default
  • Minikube URL: http://127.0.0.1:53044 (via tunnel) or http://192.168.49.2:30001 (direct)

Troubleshooting

If the API doesn't restart properly:

  1. Check pod status:

    kubectl describe pod -l app=rag-app

  2. Check events:

    kubectl get events --sort-by='.lastTimestamp' | grep rag-app

  3. Force delete and recreate:

    kubectl delete deployment rag-app-deployment
# Then redeploy using your original deployment command