I create a new class that is a subclass of multiprocessing.Process and I would like to invoke methods on this class. The methods change class members but take no arguments, and I think should work transparently. For instance, in the MWE below I create a class that inherits from Process and has a stop() function which just sets an instance member flag. When this flag is set though the run() method doesn't seem to notice a change. This all seemed to work when I was inheriting from threading.Thread, thoughts?
from queue import Empty
import multiprocessing
class Worker(multiprocessing.Process):
def __init__(self, queue):
multiprocessing.Process.__init__(self) # , daemon=True)
self.queue = queue
self.close = False
def stop(self):
self.close = True
print(self.close)
def run(self):
while (not self.close) or self.queue.qsize() > 0:
print(self.close)
print(self.queue.qsize())
for item in range(0, self.queue.qsize()):
try:
self.queue.get_nowait()
except Empty:
continue
queue = multiprocessing.Queue()
dbq = Worker(queue)
dbq.start()
queue.put("d")
dbq.stop()
dbq.join()
You have to use something like multiprocessing.Value
for synchronization between processes.
Sample code:
from queue import Empty
from ctypes import c_bool
import multiprocessing
class Worker(multiprocessing.Process):
def __init__(self, queue):
multiprocessing.Process.__init__(self) # , daemon=True)
self.queue = queue
self.close = multiprocessing.Value(c_bool, False)
def stop(self):
self.close.value = True
print(self.close)
def run(self):
while (not self.close.value) or self.queue.qsize() > 0:
print(self.close)
print(self.queue.qsize())
for item in range(0, self.queue.qsize()):
try:
self.queue.get_nowait()
except Empty:
continue
if __name__ == '__main__':
queue = multiprocessing.Queue()
dbq = Worker(queue)
dbq.start()
queue.put("d")
dbq.stop()
dbq.join()
Processes do not share memory space with their parent in the same way threads do. When a process is fork
ed it will get a new copy of the parent's memory so you can't share as easily as with threads (effectively... realistically there is copy-on-write ).
I recommend that in order to kill workers you use an synchronization primitive like Event
, because usually workers are killed together in response to something that happened.
You will end up with something like this (notice, no stop
method for workers):
from queue import Empty
import multiprocessing
class Worker(multiprocessing.Process):
# added the event to the initializing function
def __init__(self, queue, close_event):
multiprocessing.Process.__init__(self) # , daemon=True)
self.queue = queue
self.close = close_event
def run(self):
while (not self.close.is_set()) or self.queue.qsize() > 0:
print(self.close)
print(self.queue.qsize())
for item in range(0, self.queue.qsize()):
try:
self.queue.get_nowait()
except Empty:
continue
queue = multiprocessing.Queue()
# create a shared event for processes to react to
close_event = multiprocessing.Event()
# send event to all processes
dbq = Worker(queue, close_event)
dbq.start()
queue.put("d")
# set the event to stop workers
close_event.set()
dbq.join()
The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.