Question

A teammate reports that a Pod named web-app-7d9f8b6c4-x2k9j is behaving oddly. It is owned by a Deployment with 3 replicas. She runs 'kubectl delete pod web-app-7d9f8b6c4-x2k9j' to clean it up. A minute later she notices a nearly identical Pod is running again. What happened?

Accepted Answer

The ReplicaSet managed by the Deployment observed the reduced Pod count and created a replacement to maintain 3 replicas.

Answer

The kubelet automatically restarted the same Pod because its restart policy defaults to Always.

Answer

The delete command failed silently because Pods owned by a Deployment cannot be deleted directly.

Answer

The scheduler re-ran the original Pod manifest from etcd since deleted Pods are cached for one minute.