Image Pull Errors

Part of the Introduction to Kubernetes section of Coddy's Terminal journey. Lesson 23 of 45.

The most common way a Pod fails is an image that cannot be pulled: a typo in the name or a tag that does not exist. The Pod is created fine, but kubectl get pods shows ErrImagePull, and after the next attempt ImagePullBackOff: Kubernetes keeps retrying, waiting a little longer each time.

kubectl describe pod tells you exactly what went wrong. In the Events you will find a Warning such as Failed to pull image "nginx:doesnotexist" ending in not found. Because the image never started, READY stays 0/1.

The fix is to give the Pod the right image: delete it and run it again with the correct name, or for a Deployment use kubectl set image. Kubernetes never guesses what you meant, so read the event carefully.

Challenge

Beginner

Start a Pod named broken from the image nginx:doesnotexist, list the Pods to see the pull error, then describe the Pod and read the Failed event.

Try it yourself

Terminal