I have been struggling for some time to figure out how to accomplish the following: I want to delete running pod on Azure Kubernetes Service cluster on scheduled basis, so that it respawns from deployment. This is required that application re-reads configuration files stored on shared storage and shared with other application.
I have found out that Kubernetes Jobs might be handy to accomplish this, but there is some but. I cant figure how can I select corresponding pod related to my deployment as it adds random string to the deployment name, i.e
deployment-name-546fcbf44f-wckh4Using selectors to get my pod doesnt succeed as there is not such operator like LIKE
kubectl get pods --field-selector metadata.name=deployment-nameNo resources found
Looking at the official docs one way of doing this would be like so:
pods=$(kubectl get pods --selector=job-name=pi --output=jsonpath='{.items[*].metadata.name}')
echo $pods
you'd need to modify job-name to match your job name