多处理池的意外行为 map_async

unexpected behaviour of multiprocessing Pool map_async

我有一些代码可以对 python 3 应用程序中的多个文件执行相同的操作,因此看起来很适合 multiprocessing。我正在尝试使用 Pool 将工作分配给一些进程。我希望代码在进行这些计算时继续做其他事情(主要是为用户显示内容),所以我想使用 multiprocessing.Pool [=28] 的 map_async 函数=] 为此。我希望在调用它之后,代码将继续,结果将由我指定的回调处理,但这似乎并没有发生。以下代码显示了我尝试调用 map_async 的三种方式以及我看到的结果:

import multiprocessing
NUM_PROCS = 4
def func(arg_list):
    arg1 = arg_list[0]
    arg2 = arg_list[1]
    print('start func')
    print ('arg1 = {0}'.format(arg1))
    print ('arg2 = {0}'.format(arg2))
    time.sleep(1)
    result1 = arg1 * arg2
    print('end func')
    return result1

def callback(result):
    print('result is {0}'.format(result))


def error_handler(error1):
    print('error in call\n {0}'.format(error1))


def async1(arg_list1):
    # This is how my understanding of map_async suggests i should
    # call it. When I execute this, the target function func() is not called
    with multiprocessing.Pool(NUM_PROCS) as p1:
        r1 = p1.map_async(func,
                          arg_list1,
                          callback=callback,
                          error_callback=error_handler)


def async2(arg_list1):
    with multiprocessing.Pool(NUM_PROCS) as p1:
        # If I call the wait function on the result for a small
        # amount of time, then the target function func() is called
        # and executes sucessfully in 2 processes, but the callback
        # function is never called so the results are not processed
        r1 = p1.map_async(func,
                          arg_list1,
                          callback=callback,
                          error_callback=error_handler)
        r1.wait(0.1)


def async3(arg_list1):
    # if I explicitly call join on the pool, then the target function func()
    # successfully executes in 2 processes and the callback function is also
    # called, but by calling join the processing is not asynchronous any more
    # as join blocks the main process until the other processes are finished.
    with multiprocessing.Pool(NUM_PROCS) as p1:
        r1 = p1.map_async(func,
                          arg_list1,
                          callback=callback,
                          error_callback=error_handler)
        p1.close()
        p1.join()


def main():
    arg_list1 = [(5, 3), (7, 4), (-8, 10), (4, 12)]
    async3(arg_list1)

    print('pool executed successfully')


if __name__ == '__main__':
    main()

当在 main 中调用 async1async2async3 时,结果在每个函数的注释中描述。任何人都可以解释为什么不同的调用会以它们的方式运行吗?最终我想像在 async1 中那样调用 map_async,这样我就可以在工作进程繁忙时在主进程中做一些事情。我已经用 python 2.7 和 3.6 在较旧的 RH6 linux 机器和较新的 ubuntu VM 上测试了这段代码,结果相同。

发生这种情况是因为当您使用 multiprocessing.Pool 作为上下文管理器时,pool.terminate() is called when you leave the with block 会立即退出所有工作程序,而无需等待 in-progress 任务完成。

New in version 3.3: Pool objects now support the context management protocol – see Context Manager Types. __enter__() returns the pool object, and __exit__() calls terminate().

IMO 使用 terminate() 作为上下文管理器的 __exit__ 方法并不是一个很好的设计选择,因为似乎大多数人直觉地期望 close() 会被调用,这将等待 in-progress 个任务完成后再退出。不幸的是,您所能做的就是重构代码而不使用上下文管理器,或者重构代码以保证在 Pool 完成其工作之前不会离开 with 块。