Force Delete a Stuck Kubernetes Pod

Objective

Remove a pod that is stuck in Terminating state when the standard deletion grace period does not complete.

Prerequisites

• kubectl configured with access to the target cluster and namespace.
• The pod name and namespace.

Procedure

1. Identify the stuck pod:

   kubectl get pods -n <namespace>

2. Attempt a normal deletion with a short grace period:

   kubectl delete pod <pod-name> -n <namespace> --grace-period=0

3. If the pod remains, force delete it:

   kubectl delete pod <pod-name> -n <namespace> --grace-period=0 --force

4. If the pod is managed by a workload controller, scale the controller to zero and back to trigger reconciliation:

   kubectl scale deployment/<name> -n <namespace> --replicas=0
   kubectl scale deployment/<name> -n <namespace> --replicas=<original-count>

Verification

• Run kubectl get pods -n <namespace> and confirm the pod is no longer listed.
• Check the workload controller events to ensure a replacement pod is running correctly.