Saving logs from an all-purpose cluster to Volume or S3 is not consistent, because stderr, stdout, and log4j-active.log get overwritten when the cluster is restarted between minutes 01 and 59.Tested case:A job is configured to start every 20 minutes,...
