鉴于此Python程序:
# commented out code are alternatives I tried that don't work.
from multiprocessing import Process, Queue
#from multiprocessing import Process, JoinableQueue as Queue
def start_process(queue):
# queue.cancel_join_thread()
while True:
print queue.get()
if __name__ == '__main__':
queue = Queue()
# queue.cancel_join_thread()
process = Process(target=start_process, args=(queue,))
process.start()
queue.put(12)
process.join()
当我使用CTRL-C
杀死此程序时,会发生这种情况:
$> python queuetest.py
12
^CTraceback (most recent call last):
File "queuetest.py", line 19, in <module>
process.join()
File ".../python2.7/multiprocessing/process.py", line 119, in join
res = self._popen.wait(timeout)
Process Process-1:
File ".../python2.7/multiprocessing/forking.py", line 122, in wait
Traceback (most recent call last):
return self.poll(0)
File ".../python2.7/multiprocessing/forking.py", line 107, in poll
pid, sts = os.waitpid(self.pid, flag)
File ".../python2.7/multiprocessing/process.py", line 232, in _bootstrap
KeyboardInterrupt
self.run()
File ".../python2.7/multiprocessing/process.py", line 88, in run
self._target(*self._args, **self._kwargs)
File "queuetest.py", line 9, in start_process
print queue.get()
File ".../python2.7/multiprocessing/queues.py", line 91, in get
res = self._recv()
KeyboardInterrupt
..如何在信号上正确终止两个处理?
我想要实现的目标:在我的非最小程序中,第二个进程拥有一个SocketServer,需要一个额外的交互式命令行界面。
答案 0 :(得分:3)
解决方案是通过队列发送特定消息(例如,我的样本中的'exit'字符串),以终止worker(子进程)正常。当CTRL-C信号发送给所有孩子时,我们需要忽略它。 以下是示例代码:
from multiprocessing import Process, Queue
def start_process(queue):
while True:
try:
m = queue.get()
if m == 'exit':
print 'cleaning up worker...'
# add here your cleaning up code
break
else:
print m
except KeyboardInterrupt:
print 'ignore CTRL-C from worker'
if __name__ == '__main__':
queue = Queue()
process = Process(target=start_process, args=(queue,))
process.start()
queue.put(12)
try:
process.join()
except KeyboardInterrupt:
print 'wait for worker to cleanup...'
queue.put('exit')
process.join()
## or to kill anyway the worker if is not terminated after 5 seconds ...
## process.join(5)
## if process.is_alive():
## process.terminate()