🔄 Restart
Restarting an application is useful when you changed a configuration (ConfigMap or Secret) or when it froze. Kubernetes does this without taking the service down: it creates new pods and only then removes the old ones.
Restart is done with kubectl. The Nuvion Panel does not have this feature.
🚀 Before you start
- Have cluster access configured (see Cluster Access).
- Have the
web-deployDeployment from the Deploy page.
🚀 1. See the pods before
kubectl get pods -l app=web-deploy
Note the names and the AGE column.
🚀 2. Restart the Deployment
kubectl rollout restart deployment/web-deploy
Expected output: deployment.apps/web-deploy restarted
🚀 3. Follow along
kubectl rollout status deployment/web-deploy
kubectl get pods -l app=web-deploy
The pods have new names and a low AGE. The final status is
successfully rolled out.
🚀 4. Restart a single pod
When only one pod has a problem:
kubectl delete pod POD-NAME
The Deployment creates a new pod in its place.
This only works for pods created by a Deployment. A pod created on its own
(with kubectl run) does not come back after being deleted.
🚀 5. Restart other workload types
kubectl rollout restart statefulset/NAME
kubectl rollout restart daemonset/NAME
💡 When to use it after changing configuration
Environment variables coming from a ConfigMap or
Secret are only read when the pod starts. After changing
the value, run rollout restart so the application sees the change.
✅ How to know it worked
kubectl rollout status ends with successfully rolled out, and the pods have
new names and AGE.
❗ Common problems
| Symptom | Cause | Solution |
|---|---|---|
rollout status never finishes | The new pod does not become ready | See kubectl describe pod and Events |
New pods in Pending | No resources on the nodes | See Scale |
| Deleted pod does not come back | It does not belong to a Deployment | Recreate with kubectl apply |
➡️ Next step: Rollout
🧠 Questions?
Talk to our SaveinCloud technical support team!