Question

Problem: I expect child to time out and be done. but instead it times out and begins to run again.

Can anyone tell me why this program runs forever? I expect it to run one time and exit…

Here is a working program. Master threads a function to spawn a child. Works great except it ends up looping.

Here is the master:

# master.py
import multiprocessing, subprocess, sys, time

def f():
    p = subprocess.Popen(["C:\\Python32\\python.exe", "child.py"])
    # wait until child ends and check exit code
    while p.poll() == None:
        time.sleep(2)
    if p.poll() != 0:
        print("something went wrong with child.py")

# multithread a function process to launch and monitor a child
p1 = multiprocessing.Process(target = f())
p1.start()

and the child:

# child.py
import socket, sys

def main(args):
    try:
        sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
        sock.settimeout(10)
        sock.bind(('', 54324))
        data, addr = sock.recvfrom(1024) # buffer size is 1024 bytes
        print(data)
        sock.close()
        return 0
    except KeyboardInterrupt as e:
        try:
            sock.close()
            return 0
        except:
            return 0

if __name__ == "__main__":
    sys.exit(main(sys.argv))

Answer 1

The problem is that your master.py doesn’t have an if __name__ == '__main__' guard. On Windows, multiprocessing has to be able to reimport the main module in the child process, and if you don’t use this if guard, you will re-execute the multiprocessing.Process in the child (resulting in an accidental forkbomb).

To fix, simply put all of the commands in master.py in the if guard:

if __name__ == '__main__':
    # multithread a function process to launch and monitor a child
    p1 = multiprocessing.Process(target = f())
    p1.start()