Python多进程编程：深入解析与实战技巧

在写python多线程代码的时候&＃xff0c;会用到multiprocessing这个包&＃xff0c;这篇文章总结了一些这个包在多进程管理方面的一些原理和代码分析。

1. 问题一&＃xff1a;是否需要显式调用pool的close和join方法&＃xff0c;不调用的话&＃xff0c;子进程是否无法退出&＃xff1f;

首先初始化Pool的时候&＃xff0c;指定processes的个数&＃xff0c;就是pool中worker的个数&＃xff0c;pool初始化的时候&＃xff0c;会把worker以daemon&＃61;True的子进程方式启动起来。

def_repopulate_pool(self):"""Bring the number of pool processes up to the specified number,

for use after reaping workers which have exited."""

for i in range(self._processes -len(self._pool)):

w&＃61; self.Process(target&＃61;worker,

args&＃61;(self._inqueue, self._outqueue,

self._initializer,

self._initargs, self._maxtasksperchild)

)

self._pool.append(w)

w.name&＃61; w.name.replace(&＃39;Process&＃39;, &＃39;PoolWorker&＃39;)

w.daemon&＃61;True

w.start()

debug(&＃39;added worker&＃39;)

推荐在使用完pool之后&＃xff0c;用thread pool的时候调用close()和join()方法&＃xff0c;这样可以把pool中的worker都释放掉(等待子任务结束)。但是如果不显式的调用&＃xff0c;在主进程退出的时候&＃xff0c;这些子进程也会退出(原因是设置了daemon这个flag)。

def _exit_function(info&＃61;info, debug&＃61;debug, _run_finalizers&＃61;_run_finalizers,

active_children&＃61;active_children,

current_process&＃61;current_process):#NB: we hold on to references to functions in the arglist due to the

#situation described below, where this function is called after this

#module&＃39;s globals are destroyed.

global_exiting

info(&＃39;process shutting down&＃39;)

debug(&＃39;running all "atexit" finalizers with priority >&＃61; 0&＃39;)

_run_finalizers(0)if current_process() is notNone:#NB: we check if the current process is None here because if

#it&＃39;s None, any call to &＃96;&＃96;active_children()&＃96;&＃96; will throw an

#AttributeError (active_children winds up trying to get

#attributes from util._current_process). This happens in a

#variety of shutdown circumstances that are not well-understood

#because module-scope variables are not apparently supposed to

#be destroyed until after this function is called. However,

#they are indeed destroyed before this function is called. See

#issues 9775 and 15881. Also related: 4106, 9205, and 9207.

for p inactive_children():ifp._daemonic:

info(&＃39;calling terminate() for daemon %s&＃39;, p.name)

p._popen.terminate()for p inactive_children():

info(&＃39;calling join() for process %s&＃39;, p.name)

p.join()

debug(&＃39;running the remaining "atexit" finalizers&＃39;)

_run_finalizers()

主进程退出的时候&＃xff0c;会调用_exit_function, 如果看到active的children是_daemonic的就会调用其terninate方法&＃xff0c;让子进程退出。exit是通过这个调用注册的&＃xff0c;atexit.register(_exit_function)&＃xff0c;本质是利用系统的退出hook方法&＃xff0c;在退出的时候触发对应的函数。

2. 问题二&＃xff1a;如果启动之后&＃xff0c;kill -9主进程&＃xff0c;子进程会不会无法退出&＃xff1f;

如下代码是pool中worker的主代码逻辑&＃xff0c;如果kill -9主进程&＃xff0c;子进程如果没有在处理作业&＃xff0c;因为主进程退出了&＃xff0c;get()方法从queue中拿task的时候&＃xff0c;就会发生exception&＃xff0c;这样worker会退出。如果子进程正在处理任务&＃xff0c;任务结束的时候&＃xff0c;需要往queue中扔回结果&＃xff0c;因为主进程退出了&＃xff0c;所以也会exception&＃xff0c;worker一样会退出。

def worker(inqueue, outqueue, initializer&＃61;None, initargs&＃61;(), maxtasks&＃61;None):assert maxtasks is None or (type(maxtasks) &＃61;&＃61; int and maxtasks >0)

put&＃61;outqueue.put

get&＃61;inqueue.getif hasattr(inqueue, &＃39;_writer&＃39;):

inqueue._writer.close()

outqueue._reader.close()if initializer is notNone:

initializer(*initargs)

completed&＃61;0while maxtasks is None or (maxtasks and completed

task&＃61;get()except(EOFError, IOError):

debug(&＃39;worker got EOFError or IOError -- exiting&＃39;)break

if task isNone:

debug(&＃39;worker got sentinel -- exiting&＃39;)breakjob, i, func, args, kwds&＃61;tasktry:

result&＃61; (True, func(*args, **kwds))exceptException, e:

result&＃61;(False, e)try:

put((job, i, result))exceptException as e:

wrapped&＃61; MaybeEncodingError(e, result[1])

debug("Possible encoding error while sending result: %s" %(

wrapped))

put((job, i, (False, wrapped)))

completed&＃43;&＃61; 1debug(&＃39;worker exiting after %d tasks&＃39; % completed)

worker退出的时候&＃xff0c;看如下代码

## process.py

def_bootstrap(self):from . importutilglobal_current_processtry:

self._children&＃61;set()

self._counter&＃61; itertools.count(1)try:

sys.stdin.close()

sys.stdin&＃61;open(os.devnull)except(OSError, ValueError):pass_current_process&＃61;self

util._finalizer_registry.clear()

util._run_after_forkers()

util.info(&＃39;child process calling self.run()&＃39;)try:

self.run()

exitcode&＃61;0finally:

util._exit_function()

子进程run()会结束&＃xff0c;然后调用_exit_function()清理一些子进程&＃xff0c;调用_run_finalizers()结束进程。

但是如果子进程在pool的worker中跑的是长时间不退出的task&＃xff0c;那这个子进程就会无法退出&＃xff0c;一直在运行。如果task都是短作业&＃xff0c;即使主进程被kill -9&＃xff0c;子进程也会在作业跑完之后都退出。