Overview#

When you make changes to a kubernetes pod or deployment. You need to rollout these changes by hand. Which includes restarting that particular resource.

There are other options available in the documentation for the command. But, I generally run into what we’re going to see here. If I run into more. I’ll probably create an article catered to that scenario.

Restarting Kubernetes Deployments#

Restart a specified kubernetes deployment within a certain namespace. In this case it’s the lab namespace.

kubectl -n lab rollout restart deployment/lab-app

Restart a specified kubernetest deployment within the default namespace.

kubectl rollout restart deployment/lab-app

To restart all deployments within a specific namespace. You can use the following command.

kubectl rollout restart deployment -n lab

If this is successful. You should see the following output stating the deployement has been restarted.

deployment.apps/lab-app restarted

Once this is finished. You can get your pod to make sure everything looks as expected. Check the age of the pod and whether there are any duplicates.

NAME                                                 READY   STATUS    RESTARTS   AGE
lab-app-343432k344-flhcr                             1/1     Running   0          7s

I’ve seen with prometheus server pods keep an old exporter around for whatever reason. Though, don’t quote me on that. This could have been because I was too impatient and didn’t wait long enough for the pod to go away.

Conclusion#

Discussed how to restart a kubernetes deployment using the rollout restart commands in kubectl. Also reviewed ways to confirm everything is good to go.