Python Process Pool non-daemonic?

Question

Would it be possible to create a python Pool that is non-daemonic? I want a pool to be able to call a function that has another pool inside.

I want this because deamon processes cannot create process. Specifically, it will cause the error:

AssertionError: daemonic processes are not allowed to have children

For example, consider the scenario where function_a has a pool which runs function_b which has a pool which runs function_c. This function chain will fail, because function_b is being run in a daemon process, and daemon processes cannot create processes.

Solution 1

The multiprocessing.pool.Pool class creates the worker processes in its __init__ method, makes them daemonic and starts them, and it is not possible to re-set their daemon attribute to False before they are started (and afterwards it's not allowed anymore). But you can create your own sub-class of multiprocesing.pool.Pool (multiprocessing.Pool is just a wrapper function) and substitute your own multiprocessing.Process sub-class, which is always non-daemonic, to be used for the worker processes.

Here's a full example of how to do this. The important parts are the two classes NoDaemonProcess and MyPool at the top and to call pool.close() and pool.join() on your MyPool instance at the end.

#!/usr/bin/env python
# -*- coding: UTF-8 -*-

import multiprocessing
# We must import this explicitly, it is not imported by the top-level
# multiprocessing module.
import multiprocessing.pool
import time

from random import randint


class NoDaemonProcess(multiprocessing.Process):
    # make 'daemon' attribute always return False
    def _get_daemon(self):
        return False
    def _set_daemon(self, value):
        pass
    daemon = property(_get_daemon, _set_daemon)

# We sub-class multiprocessing.pool.Pool instead of multiprocessing.Pool
# because the latter is only a wrapper function, not a proper class.
class MyPool(multiprocessing.pool.Pool):
    Process = NoDaemonProcess

def sleepawhile(t):
    print("Sleeping %i seconds..." % t)
    time.sleep(t)
    return t

def work(num_procs):
    print("Creating %i (daemon) workers and jobs in child." % num_procs)
    pool = multiprocessing.Pool(num_procs)

    result = pool.map(sleepawhile,
        [randint(1, 5) for x in range(num_procs)])

    # The following is not really needed, since the (daemon) workers of the
    # child's pool are killed when the child is terminated, but it's good
    # practice to cleanup after ourselves anyway.
    pool.close()
    pool.join()
    return result

def test():
    print("Creating 5 (non-daemon) workers and jobs in main process.")
    pool = MyPool(5)

    result = pool.map(work, [randint(1, 5) for x in range(5)])

    pool.close()
    pool.join()
    print(result)

if __name__ == '__main__':
    test()

Solution 2

I had the necessity to employ a non-daemonic pool in Python 3.7 and ended up adapting the code posted in the accepted answer. Below there's the snippet that creates the non-daemonic pool:

import multiprocessing.pool

class NoDaemonProcess(multiprocessing.Process):
    @property
    def daemon(self):
        return False

    @daemon.setter
    def daemon(self, value):
        pass


class NoDaemonContext(type(multiprocessing.get_context())):
    Process = NoDaemonProcess

# We sub-class multiprocessing.pool.Pool instead of multiprocessing.Pool
# because the latter is only a wrapper function, not a proper class.
class NestablePool(multiprocessing.pool.Pool):
    def __init__(self, *args, **kwargs):
        kwargs['context'] = NoDaemonContext()
        super(NestablePool, self).__init__(*args, **kwargs)

As the current implementation of multiprocessing has been extensively refactored to be based on contexts, we need to provide a NoDaemonContext class that has our NoDaemonProcess as attribute. NestablePool will then use that context instead of the default one.

That said, I should warn that there are at least two caveats to this approach:

1. It still depends on implementation details of the multiprocessing package, and could therefore break at any time.
2. There are valid reasons why multiprocessing made it so hard to use non-daemonic processes, many of which are explained here. The most compelling in my opinion is:

As for allowing children threads to spawn off children of its own using subprocess runs the risk of creating a little army of zombie 'grandchildren' if either the parent or child threads terminate before the subprocess completes and returns.

Solution 3

The multiprocessing module has a nice interface to use pools with processes or threads. Depending on your current use case, you might consider using multiprocessing.pool.ThreadPool for your outer Pool, which will result in threads (that allow to spawn processes from within) as opposed to processes.

It might be limited by the GIL, but in my particular case (I tested both), the startup time for the processes from the outer Pool as created here far outweighed the solution with ThreadPool.

---

It's really easy to swap Processes for Threads. Read more about how to use a ThreadPool solution here or here.

Solution 4

As of Python 3.8, concurrent.futures.ProcessPoolExecutor doesn't have this limitation. It can have a nested process pool with no problem at all:

from concurrent.futures import ProcessPoolExecutor as Pool
from itertools import repeat
from multiprocessing import current_process
import time

def pid():
    return current_process().pid

def _square(i):  # Runs in inner_pool
    square = i ** 2
    time.sleep(i / 10)
    print(f'{pid()=} {i=} {square=}')
    return square

def _sum_squares(i, j):  # Runs in outer_pool
    with Pool(max_workers=2) as inner_pool:
        squares = inner_pool.map(_square, (i, j))
    sum_squares = sum(squares)
    time.sleep(sum_squares ** .5)
    print(f'{pid()=}, {i=}, {j=} {sum_squares=}')
    return sum_squares

def main():
    with Pool(max_workers=3) as outer_pool:
        for sum_squares in outer_pool.map(_sum_squares, range(5), repeat(3)):
            print(f'{pid()=} {sum_squares=}')

if __name__ == "__main__":
    main()

The above demonstration code was tested with Python 3.8.

A limitation of ProcessPoolExecutor, however, is that it doesn't have maxtasksperchild. If you need this, consider the answer by Massimiliano instead.

Credit: answer by jfs

Solution 5

On some Python versions replacing standard Pool to custom can raise error: AssertionError: group argument must be None for now.

Here I found a solution that can help:

class NoDaemonProcess(multiprocessing.Process):
    # make 'daemon' attribute always return False
    @property
    def daemon(self):
        return False

    @daemon.setter
    def daemon(self, val):
        pass


class NoDaemonProcessPool(multiprocessing.pool.Pool):

    def Process(self, *args, **kwds):
        proc = super(NoDaemonProcessPool, self).Process(*args, **kwds)
        proc.__class__ = NoDaemonProcess

        return proc

View more solutions

Comments

[Admin]

The above code seems to be hanging for me. Specifically it appears to hang at pool.close() inside work(). Is there anything i am missing ?

[Chris Arndt]

I just tested my code again with Python 2.7/3.2 (after fixing the "print" lines) on Linux and Python 2.6/2.7/3.2 OS X. Linux and Python 2.7/3.2 on OS X works fine but the code does indeed hang with Python 2.6 on OS X (Lion). This seems to be a bug in the multiprocessing module, which got fixed, but I haven't actually checked the bug tracker.

[Mike Vella]

This should really be fixed in the multiprocessing module (an option for non-daemonic workers should be available). Does anyone know who maintains it?

[frmdstryr]

Thanks! On windows you also need to call multiprocessing.freeze_support()

[Chris Lucian]

Nice work. If anyone is getting memory leak with this try using "with closing(MyPool(processes=num_cpu)) as pool:" to dispose of the pool properly

[max]

What's the disadvantages of using MyPool instead of the default Pool? In other words, in exchange for the flexibility of starting child processes, what costs do I pay? (If there were no costs, presumably the standard Pool would have used non-daemonic processes).

[Philippe Ombredanne]

@ChrisArndt what would be the license for your code beside the standard CC-CY-SA from SO? MIT?

[Chris Arndt]

@PhilippeOmbredanne: Yes, MIT is fine. Most code I publish, which doesn't have to be GPL, is under MIT license.

[Philippe Ombredanne]

@ChrisArndt Thanks... I am going to put to good use in github.com/nexB/scancode-toolkit !

[Chris Arndt]

@machen Yes, unfortunately that's true. In Python 3.6 the Pool class has been extensively refactored, so Process isn't a simple attribute anymore, but a method, which returns the process instance it gets from a context. I tried overwriting this method to return a NoDaemonPool instance, but this results in the exception AssertionError: daemonic processes are not allowed to have children when the Pool is used.

[Chris Arndt]

Just tested it again with Python 3.7.3, and, to my surprise, this still works. Rather accidentally, though. Overriding multiprocessing.pool.Pool.Process circumvents the whole new context stuff, and I'm not sure what side-effects this has.

[gbmhunter]

Confirmed that it works for me in Python 3.6.5 (perhaps the answer was edited to support >=v3.6 after the comment It doesn't work with python3.6?

[A_A]

Regarding the caveat: My use case is parallelising tasks, but the grand-children return information to their parents that in turn return information to their parents after doing some required local processing. Consequently, every level / branch has an explicit wait for all its leafs. Does the caveat still apply if you explicitly have to wait for spawned processes to finish?

[kadee]

It doesn't work in Python 3.6.7, most probably due to this commit: github.com/python/cpython/commit/…. That commit was reverted already in Python 3.6.8 (github.com/python/cpython/pull/10969). So, I guess, it's only in Python 3.6.7 that this solution is not working?

[trance_dude]

Thanks - this helped me a lot - great use of threading here (to spawn processes which actually perform well)

[Max Wong]

I just used the answer, but I received this AttributeError: module 'multiprocessing' has no attribute 'pool' error. Why?

[Chris Arndt]

@YanqiHuang Don't forget import multiprocessing.pool

[Max Wong]

@ChrisArndt I tried import multiprocessing and then used multiprocessing.pool.Pool as the parent class for the MyPool class. Will there be any difference between the two? Thanks

[Chris Arndt]

@YanqiHuang No, that won't work. You need to import multiprocessing.pool, exactly as I wrote.

[Max Wong]

@ChrisArndt Oh... I see Thanks a lot :)

[Max Wong]

@ChrisArndt Could you also explain why we need the way as you suggested. Thanks a million. This is just our of my curiosity

[DreamFlasher]

This is now clearly the best solution, as it requires minimal changes.

[abanana]

For people looking for a practical solution that probably applies to their situation, this is the one.

[raphael]

works perfectly! ... as a side-note using a child- multiprocessing.Pool inside a ProcessPoolExecutor.Pool is also possible!

[Radio Controlled]

Would you bother adding how to use this instead of multiprocessing.pool?

[Radio Controlled]

"You can now use multiprocessing.Pool and NestablePool interchangeably".

[Roy Shilkrot]

Unfortunately this doesn't work for me, still getting daemonic processes are not allowed to have children

[Asclepius]

@RoyShilkrot Which version of Python are you using exactly?

[Roy Shilkrot]

python 3.7. The problem was this was run from Celery, and I had to use import billiard as multiprocessing and use their Pool.

[Asclepius]

@RoyShilkrot Noted. The current latest version of Celery claims to support Python 3.8 which is the version of Python that the answer was tested with.

[wim]

Users choosing a process pool are presumably CPU-bound and/or need cancellable tasks, so threads are not an option. This doesn't really answer the question.

[Roel]

This won't work if your script is within celery though. Using celery==4.3.0 and python==3.7

[Opps_0]

@Acumenus would you mind checking my question (stackoverflow.com/questions/68305077/…)

[Asclepius]

@Opps_0 What's to check. Read the docs of concurrent.futures.ProcessPoolExecutor, get it to work with something very simple, and then with your actual task.