python多处理：睡眠语句被杀死？

Question

我有以下代码：

import multiprocessing, datetime

def Unit_Task_Function(Argument):

    print(f"Unit of work {Argument} starting {datetime.datetime.now().strftime('%Y-%m-%d : %H-%M-%S')}")
    sleep(2*random.random())
    print(f"Unit of work {Argument} ending {datetime.datetime.now().strftime('%Y-%m-%d : %H-%M-%S')}")    

if __name__ == "__main__":  # Allows for the safe importing of the main module
    __spec__ = "ModuleSpec(name='builtins', loader=<class '_frozen_importlib.BuiltinImporter'>)"

    print(f"There are {multiprocessing.cpu_count()} CPUs on this machine")

    max_para_processes_at_any_time = 5
    pool = multiprocessing.Pool(max_para_processes_at_any_time)


    iterable_arguments = ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'M', 'N']

    results = pool.map_async(Unit_Task_Function, iterable_arguments)

    pool.close()
    pool.join()

输出是：

In [18]: run code.py
There are 8 CPUs on this machine
Unit of work A starting 2020-01-27 : 12-26-02
Unit of work B starting 2020-01-27 : 12-26-02
Unit of work C starting 2020-01-27 : 12-26-02
Unit of work D starting 2020-01-27 : 12-26-02
Unit of work E starting 2020-01-27 : 12-26-02
Unit of work F starting 2020-01-27 : 12-26-02
Unit of work G starting 2020-01-27 : 12-26-02
Unit of work H starting 2020-01-27 : 12-26-02
Unit of work M starting 2020-01-27 : 12-26-02
Unit of work N starting 2020-01-27 : 12-26-02

为什么第二个打印语句从未完成？ 这与所谓的“守护进程”进程有关吗？ 如何重写代码以使其工作？

Answer 1

你缺少一些进口。 这会导致工作人员出现导入错误，这是您看不到的。 导入这些：

import random

from time import sleep

你这个输出：

There are 8 CPUs on this machine
Unit of work A starting 2020-01-27 : 18-48-21
Unit of work B starting 2020-01-27 : 18-48-21
Unit of work C starting 2020-01-27 : 18-48-21
Unit of work D starting 2020-01-27 : 18-48-21
Unit of work E starting 2020-01-27 : 18-48-21
Unit of work E ending 2020-01-27 : 18-48-21
Unit of work F starting 2020-01-27 : 18-48-21
Unit of work A ending 2020-01-27 : 18-48-22
Unit of work G starting 2020-01-27 : 18-48-22
Unit of work C ending 2020-01-27 : 18-48-22
Unit of work H starting 2020-01-27 : 18-48-22
Unit of work D ending 2020-01-27 : 18-48-22
Unit of work M starting 2020-01-27 : 18-48-22
Unit of work G ending 2020-01-27 : 18-48-22
Unit of work N starting 2020-01-27 : 18-48-22
Unit of work B ending 2020-01-27 : 18-48-22
Unit of work M ending 2020-01-27 : 18-48-23
Unit of work F ending 2020-01-27 : 18-48-23
Unit of work N ending 2020-01-27 : 18-48-24
Unit of work H ending 2020-01-27 : 18-48-24

提示：在切换到pool.map_async()之前，首先使用pool.map()查看错误消息。

或者，提供一个重新引发异常的错误回调函数：

def on_error(err):
        raise err

results = pool.map_async(Unit_Task_Function, 
                         iterable_arguments,
                         error_callback=on_error)