Detailed Debug Msg w/ KubeCluster

We got the distributed scheduler/worker/dashboard all working well as either pull docker image or standalone on local Mac box. But we cannot get it work using KubeCluster(pod_spec, …) on single machine where pod_spec is defined ahead w/o error msg. We can see pods got started OK from K8s side but shut down very soon. We check the returned future from submission (future = client.submit()) as “error” but no clue on what went wrong. Anyone runs into similar problem? Or any suggestion in getting more detailed error msg ? Thanks

@dask_users Welcome! Apologies for the delay in response.

Have you resolved this? If not, could you please share some more details about your cluster and Kubernetes configuration?

In general, something in this doc might help. You can also dump your cluster state, which can help with debugging. :smile: