Here’s a really quick tip about an annoying Kubernetes issue I’ve encountered from time to time while updating my deployments.
Sometimes, for obscure reasons, pods get stuck in the “Terminating” state. There’s an open issue about this on Github.
It can apparently happen because of a lack of memory, timeouts, issues with volumes, etc. Since the issue is not resolved, there are only workarounds at this time.
The one that has worked for me is the following:
kubectl delete pods <pod> --grace-period=0 --force
This kills the problematic pods (pretty violently). Personally I only do this on my development workstation, so it’s all right. But be careful if you use this on a production system ;-)