Documentation improvement for multiprocessing: a script which spawns processes on import will cause system failure

[stscijgbot-jp]

Issue JP-2980 was created on JIRA by Tyler Pauly:

A help desk ticket (INC0182968) has exposed a weakness in our documentation regarding the use of parallel processing in the pipeline.

If multiprocessing is active, when running from a script (i.e. python multiproc_detector1.py), importing that script must not spawn more processes. This is typically prevented by placing process-spawning code inside the if __name__ == '__main__': block.

Users are not warned of this requirement in the documentation; if not followed, this will lead to a system crash as the code will attempt to spawn an ever-increasing number of processes.

[stscijgbot-jp]

Comment by Tyler Pauly on JIRA:

From standup discussion today, this ticket links to the "nested multiprocessing" issue described by a user: https://jira.stsci.edu/browse/JP-3586

[stscijgbot-jp]

Comment by Maria Pena-Guerrero on JIRA:

Documentation included by: #8408

[stscijgbot-jp]

Comment by Howard Bushouse on JIRA:

Fixed by #8408