由于pyton process不共享内存，如果需要共享数据，需要采用它只有的Value或者Manager方式

#!/usr/env python
import multiprocessing as mp
from multiprocessing import Value, Array
import threading
import time
import pdb
 
def init(args):
    ''' store the counter for later use '''
    global counter
    counter = args

def run():
    global counter
    for i in range(0, 3):
        time.sleep(1);
        with counter.get_lock():
            counter.value += 1
        print counter.value

if __name__ == "__main__":
    counter = Value('i', 0)
    pool = mp.Pool(processes = 5,initializer = init, initargs = (counter, ))
    print "I am set counter:%s" % (counter.value)
    for i in range(0, 10):
        pdb.set_trace()
        res = pool.apply_async(run, ())
    pool.close()
    pool.join()
    print "I am set counter:%s" % (counter.value)