多线程介绍和多线程模块-lock-互斥锁
多线程介绍和多线程模块
线程的特点:
线程的生命周期
开始
运行
结束
线程的退出:
进程执行完成
线程的退出方法
python的系统推出
模块函数
start_new_thread(func, args) #(func,(name,i))
allocate_lock()
exit()
[root@133 managehosts]# vim thread01.py #!/usr/bin/env python#encoding:utf8import threadimport timedef func(name , i): for n in xrange(i): print name, n time.sleep(1)thread.start_new_thread(func,('声音', 3)) #(func,(name , i))thread.start_new_thread(func,('画面', 3))time.sleep(3)[root@133 managehosts]# python thread01.py 声音 0画面 0声音 1画面 1声音 2画面 2
LockType对象方法
lock= thread.allocate_lock() 生成锁对象
lock.acquire() 加锁
lock.locked() 查看锁状态
lock.release() 多线程调度释放锁,使用前线程必须已经获得锁定,否则将抛出异常
[root@133 managehosts]# vim thread01.py #!/usr/bin/env python#encoding:utf8import threadimport timedef func(name , i, l): for n in xrange(i): print name, n time.sleep(1) l.release() #release解锁lock = thread.allocate_lock() #申请一把锁lock.acquire() #获取并使用这把锁thread.start_new_thread(func,('声音', 5, lock))thread.start_new_thread(func,('画面', 5, lock))while lock.locked(): #这是主进程,lock.locked=True,执行pass,主进程一直处于锁定状态,线程执行,直到线程执行完毕,释放锁,lock.locked=False passprint lock.locked() #解锁后,值为Falseprint "Exit main process"[root@133 managehosts]# python thread01.py 声音 画面 00画面 1声音 1画面 2声音 2画面 3声音 3画面 4声音 4FalseExit main process[root@133 managehosts]# vim thread02.py #!/usr/bin/env pythonimport threadimport timedef printworld(): for i in range(5): if w_lock.acquire(): print 'world',time.ctime() h_lock.release()h_lock = thread.allocate_lock()w_lock = thread.allocate_lock()thread.start_new_thread(printworld, ())w_lock.acquire()for i in range(5): if h_lock.acquire(): print 'hello', w_lock.release()time.sleep(1)[root@133 managehosts]# python thread02.py hello world Fri Feb 10 20:26:51 2017hello world Fri Feb 10 20:26:51 2017hello world Fri Feb 10 20:26:51 2017hello world Fri Feb 10 20:26:51 2017hello world Fri Feb 10 20:26:51 2017[root@133 managehosts]# vim thread02.py #!/usr/bin/env pythonimport threadimport timedef printworld(): for i in range(5): if w_lock.acquire(): #申请world锁,if True为真 print 'world',time.ctime() #打印world h_lock.release() #释放hello锁h_lock = thread.allocate_lock()w_lock = thread.allocate_lock()thread.start_new_thread(printworld, ()) #开启的线程,和主进程并列执行。w_lock.acquire() #获得world锁for i in range(5): if h_lock.acquire(): #申请hello锁,if True为真 print 'hello', #打印hello w_lock.release() #释放world锁while h_lock.locked(): #这是主进程,直到hello打印5次后才释放。 pass [root@133 managehosts]# python thread02.py hello world Fri Feb 10 20:29:48 2017hello world Fri Feb 10 20:29:48 2017hello world Fri Feb 10 20:29:48 2017hello world Fri Feb 10 20:29:48 2017hello world Fri Feb 10 20:29:48 2017
申请了锁之后,会执行thread.start_new_thread(func,('声音',5,lock))和thread.start_new_thread(func,('画面',5,lock))这两行代码的。
执行他们的时候,同时也会执行主进程里的while的。主进程与这两个线程是同时执行的。为了不让线程退出,所以在主进程里有while来判断锁是不是已经释放了,如果是释放了,说明线程执行完了。
使用主线程控制,线程打印hello和world
[root@133 managehosts]# vim thread03.py +22#!/usr/bin/env pythonimport threadimport timedef hello(): for i in xrange(5): h_lock.acquire() print 'hello', w_lock.release()def world(): for i in xrange(5): w_lock.acquire() print 'world',time.ctime() h_lock.release() lock.release()#释放主线程的的锁lock = thread.allocate_lock() #为主进程申请锁lock.acquire() #主进程获得锁h_lock = thread.allocate_lock() #为hello线程申请一把锁w_lock = thread.allocate_lock() #为world线程申请一把锁w_lock.acquire() #world线程获取锁thread.start_new_thread(hello,())#启动线程hellothread.start_new_thread(world,())#启动线程world#time.sleep(1)while lock.locked(): #当主线程的锁的状态为true时 pass #什么都不做 #hello 和 world 的两个线程是同时运行的 [root@133 managehosts]# python thread03.py hello world Tue Feb 14 14:59:11 2017hello world Tue Feb 14 14:59:11 2017hello world Tue Feb 14 14:59:11 2017hello world Tue Feb 14 14:59:11 2017hello world Tue Feb 14 14:59:11 2017
从脚本的执行顺序,hello()这个函数先执行,输出hello,
然后把world的锁释放(w_lock.release())。
world()函数获得锁,输出world,然后把hello的锁释放。
必须释放对方的锁,对方才能获得锁(h_lock.acquire()),如果对方的锁没有被释放,那么对方再想获得锁是不成功的,会一直处于阻塞状态,直到锁被释放。
hello()函数第一次获得了锁资源(h_lock.acquire()),打印了hello,接着释放了world的锁资源(w_lock.release())。
world()函数获得了锁资源(w_lock.acquire()),打印了world,接着释放了hello的锁资源(h_lock.release())。
由于hello的锁资源被释放了,hello函数里又获得了锁资源,接着第二次打印hello。
也就是说哪个函数或得了锁资源,就会打印相应的信息。
锁资源只有被释放了,下一次才能获得,即已经获得了锁资源,不能再次获得,所以在hello函数里释放world的锁资源,在world函数里才能获得锁资源。
这样反反复复,就打印了hello world。
threading
threading不需要进程的控制来控制线程
threading.Thread:类
成员方法:
start() 启动线程
run() 可以重写
join() 阻塞
getName()
setName()
isDaemon()判断线程是否随主线程一起结束
setDaemon设置线程与主线程一起结束
In [1]: import threadingIn [2]: help(threading.Thread)Help on class Thread in module threading:class Thread(_Verbose) | Method resolution order: | Thread | _Verbose | __builtin__.object | | Methods defined here: | | __init__(self, group=None, target=None, name=None, args=(), kwargs=None, verbose=None)
[root@133 managehosts]# vim threading01.py#!/usr/bin/env pythonimport threadingimport timedef func(): print 'hello',time.ctime() time.sleep(1)if __name__ == '__main__': for i in xrange(10): t = threading.Thread(target=func, args =()) t.start()print 'complete' [root@133 managehosts]# python threading01.pyhello Tue Feb 14 15:42:10 2017 #瞬间打印10行hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017complete
如果添加t.join,则会阻塞,每秒打印一次,上一个阻塞执行完才会执行下一个
def func(): print 'hello',time.ctime() time.sleep(1)if __name__ == '__main__': for i in xrange(10): t = threading.Thread(target=func, args =()) t.start() t.join()[root@133 managehosts]# python threading01.py hello Tue Feb 14 15:42:10 2017 #每隔一秒打印一次hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017hello Tue Feb 14 15:42:10 2017complete
[root@133 managehosts]# vim threading02.py #!/usr/bin/env python#encoding:utf8import threadingimport timedef func(name, i): for n in xrange(i): print name, n time.sleep(1)t1 = threading.Thread(target=func, args=('声音',3))t2 = threading.Thread(target=func, args=('画面',3))t1.start()t2.start()[root@133 managehosts]# python threading02.py 声音 0画面 0声音 1画面 1声音 2画面 2如果添加一个t1.join(),则会先打印t1,然后再打印t2t1.start()t1.join()t2.start()结果变为:[root@133 managehosts]# python threading02.py 声音 0声音 1声音 2画面 0画面 1画面 2
t1.setDaemon(True) 设置t1线程和主进程一起退出
[root@133 managehosts]# vim threading02.py #!/usr/bin/env python#encoding:utf8import threadingimport timedef func(name, i): for n in xrange(i): print name, n time.sleep(1)t1 = threading.Thread(target=func, args=('声音',3))t2 = threading.Thread(target=func, args=('画面',3))t1.setDaemon(True) #设置t1线程和主进程一起退出t1.start() #开启t1线程,打印一次就会退出,因为主进程已经退出t2.setDaemon(True)t2.start()[root@133 managehosts]# python threading02.py 声音 0画面 0
互斥锁
[root@133 managehosts]# pwd/opt/python/managehosts[root@133 managehosts]# cat threading07.py#!/usr/bin/env pythonimport threadingimport timedef hello(): for i in xrange(5): h_lock.acquire() print 'hello', w_lock.release()def world(): for i in xrange(5): w_lock.acquire() print 'world',time.ctime() h_lock.release()h_lock = threading.Lock()w_lock = threading.Lock()w_lock.acquire()t1 = threading.Thread(target=hello, args=())t2 = threading.Thread(target=world, args=())t1.start()t2.start()[root@133 managehosts]# python threading07.pyhello world Wed Mar 29 11:51:59 2017hello world Wed Mar 29 11:51:59 2017hello world Wed Mar 29 11:51:59 2017hello world Wed Mar 29 11:51:59 2017hello world Wed Mar 29 11:51:59 2017
#!/usr/bin/pythonimport threadimport timedef hello(): for i in xrange(5): h_lock.acquire() print 'hello', w_lock.release()def world(): for i in xrange(5): w_lock.acquire() print 'world' h_lock.release() lock.release()lock = thread.allocate_lock() lock.acquire() h_lock = thread.allocate_lock()w_lock = thread.allocate_lock()w_lock.acquire()thread.start_new_thread(hello, ())thread.start_new_thread(world, ())while lock.locked(): pass
程序执行开两个线程,即hello和world。
执行hello线程时,只有3步:
先申请h_lock锁,即h_lock.acquire()
然后输出hello,即print 'hello',
最后释放w_lock锁,即w_lock.release()
执行world线程时也是3步:
申请w_lock锁,即w_lock.acquire()
输出world,即print 'world'
释放h_lock锁,即h_lock.release()
线程是同时执行的,为什么hello和world会交替输出,没有先输出world?
在主线程里,申请了w_lock锁,即26行的w_lock.acquire(),所以在world()函数里的w_lock.acquire()一直
处于阻塞状态,意味着print 'world'这一行是不会执行的,直到hello()函数里把w_lock释放,即w_lock.release()
所以在每个函数里都有释放对方的锁,这样对方就把自己要输出的内容执行了。
最后world函数里的循环执行完了,释放lock这个锁,这样这个主进程也结束了。
[root@133 managehosts]# cat threading08.py#!/usr/bin/env pythonimport threadingimport timeclass MyThread(threading.Thread): def __init__(self): threading.Thread.__init__(self) def run(self): global counter time.sleep(1) mutex.acquire() #获得锁 counter +=1 print "I am %s, set counter:%s" %(self.name,counter) mutex.release() #释放锁if __name__ == '__main__': counter = 0 mutex = threading.Lock() #分配锁 for i in xrange(300): t = MyThread() t.start() [root@133 managehosts]# python threading08.pyI am Thread-1, set counter:1I am Thread-2, set counter:2I am Thread-3, set counter:3I am Thread-4, set counter:4I am Thread-5, set counter:5I am Thread-6, set counter:6I am Thread-7, set counter:7I am Thread-8, set counter:8I am Thread-9, set counter:9I am Thread-10, set counter:10I am Thread-11, set counter:11I am Thread-12, set counter:12I am Thread-13, set counter:13I am Thread-14, set counter:14I am Thread-15, set counter:15I am Thread-284, set counter:282I am Thread-283, set counter:283I am Thread-282, set counter:284I am Thread-285, set counter:285I am Thread-287, set counter:286I am Thread-286, set counter:287I am Thread-288, set counter:288I am Thread-289, set counter:289I am Thread-290, set counter:290I am Thread-293, set counter:291I am Thread-292, set counter:292I am Thread-291, set counter:293I am Thread-295, set counter:294I am Thread-296, set counter:295I am Thread-297, set counter:296I am Thread-294, set counter:297I am Thread-298, set counter:298I am Thread-299, set counter:299I am Thread-300, set counter:300