I have noticed on our local install, that the job scripts as well as stdout/err in our cluster_files_directory is not getting cleaned up. When I track the jobs in our runner logs, for a sampling of the job scripts left over in the cluster_files_directory, the jobs are completing out successfully and there are no errors in the log. I have not tried to check if all scripts are not getting cleaned, or just some … but by the volume of them it looks like all of them may not be getting removed/cleaned up after the job completes out.