Modules own many orphaned sockets file descriptors due to the fork() of multiprocessing

[AlyaGomaa]

To reproduce:

• strace -f -p pid_of_a_slips_module
• ls -l /proc/<pid_of_a_slips_module>/fd/

you will find modules looping through FDs, listening, then timing out, etc.
this leads to unneccessary CPU usage

Cause:
When creating a redis instance in slips main.py before forking the modules, the fork copies all FDs to the children

Solution: Use a multiprocessing context that uses spawn instead of fork

import multiprocessing as mp

mp.set_start_method("spawn") # instead of fork