Batch processing failure with Detector1 multiprocessing

[stscijgbot-jp]

Issue JP-3726 was created on JIRA by Melanie Clarke:

A user reports that batch processing around 200 files through Detector1 in the same Python process eventually fails with a resource warning.  They note that it occurs when multiprocessing arguments are passed for jump and ramp fitting, but the processing completes okay when single-threaded.

Possibly related: multiprocessing in jump and ramp fitting often reports the following warning after the Python process closes: "leaked semaphore objects to clean up at shutdown".  Log from the user reported 9886 leaked semaphore objects.

Helpdesk ticket is: INC0203480

The workaround at the moment is to reduce fewer files per Python process, or to turn off multiprocessing in jump and ramp fit.