简介: zeromq中间件,他是一个轻量级的消息中间件,传说是世界上最快的消息中间件,为什么这么说呢? 因为一般的消息中间件都需要启动消息服务器,但是zeromq这厮尽然没有消息服务器,他压根没有消息中间件的架子,但是这并不能掩盖他的强大。 通过和activemq,rabbitmq对比,显然功能上没有前两者这么强大,他不支持消息的持久化,但是有消息copy功能,他也不支持崩溃恢复,而且由于他太快了,可能客户端还没启动,服务端的消息就已经发出去了,这个就容易丢消息了,但是zeromq自由他的办法,就先说这么多了。先来看看怎么在python中引入这个强大的利器。 我自己之所以,学习体会一下,主要原因,是想在练习过程中体会其中的应用原理及逻辑,最好是能感知到其中的设计思想,为以后,自己做东西积攒点经验. 另外最近也比较关注自动化运维的一些东西.网上说saltstack本身就用的zeromq做消息队列.所以更引起了我的兴趣. 安装: 我的操作系统是ubuntu 14.04的 python zeromq 环境安装参考这里的官网 下面测试: 一,C/S模式: server 端代码: #!/usr/bin/env python # coding:utf8 #author: wangqiankun@lashou-inc.com import zmq #调用zmq相关类方法,邦定端口 context = zmq.Context() socket = context.socket(zmq.REP) socket.bind('tcp://*:10001') while True: #循环接受客户端发来的消息 msg = socket.recv() print "Msg info:%s" %msg #向客户端服务器发端需要执行的命令 cmd_info = raw_input("client cmd info:").strip() socket.send(cmd_info) socket.close() client 端代码: import zmq import time import commands context = zmq.Context() socket = context.socket(zmq.REQ) socket.connect('tcp://127.0.0.1:10001') def execut_cmd(cmd): s,v = commands.getstatusoutput(cmd) return v while True: #获取当前时间 now_time = time.strftime("%Y-%m-%d %H:%M:%S",time.localtime()) socket.send("now time info:[%s] request execution command:' ',%s"%(now_time,result)) recov_msg = socket.recv() #调用execut_cmd函数,执行服务器发过来的命令 result = execut_cmd(recov_msg) print recov_msg,' ',result, time.sleep(1) #print "now time info:%s cmd status:[%s],result:[%s]" %(now_time,s,v) continue socket.close() 注意:此模式是经典的应答模式,不能同时send多个数据, 这种模式说是主要用于远程调用和任务分配,但我愚笨,还是理解不透.后面有时间,再回过来好好看看, 测试: req端 # python zmq-server-cs-v01.py rep端 # python zmq-client-cs-v01.py 二,发布订阅模式(pub/sub) pub 发布端代码如下: #!/usr/bin/env python # coding:utf8 #author: wangqiankun@lashou-inc.com import itertools import sys,time,zmq def main(): if len(sys.argv) != 2: print 'Usage: publisher' sys.exit(1) bind_to = sys.argv[1] all_topics = ['sports.general','sports.football','sports.basketball','stocks.general','stocks.GOOG','stocks.AAPL','weather'] ctx = zmq.Context() s = ctx.socket(zmq.PUB) s.bind(bind_to) print "Starting broadcast on topics:" print "%s" %all_topics print "Hit Ctrl-c to stop broadcasting." print "waiting so subscriber sockets can connect...." print time.sleep(1) msg_counter = itertools.count() try: for topic in itertools.cycle(all_topics): msg_body = str(msg_counter.next()) #print msg_body, print 'Topic:%s,msg:%s' %(topic,msg_body) s.send_multipart([topic,msg_body]) #s.send_pyobj([topic,msg_body]) time.sleep(0.1) except KeyboardInterrupt: pass print "Wating for message queues to flush" time.sleep(0.5) s.close() print "Done" if __name__ == "__main__": main() sub 端代码: #!/usr/bin/env python # coding:utf8 #author: wangqiankun@lashou-inc.com import zmq import time,sys def main(): if len(sys.argv) < 2: print "Usage: subscriber [topic topic]" sys.exit(1) connect_to = sys.argv[1] topics = sys.argv[2:] ctx = zmq.Context() s = ctx.socket(zmq.SUB) s.connect(connect_to) #manage subscriptions if not topics: print "Receiving messages on ALL topics...." s.setsockopt(zmq.SUBSCRIBE,'') else: print "Receiving messages on topics: %s..." %topics for t in topics: s.setsockopt(zmq.SUBSCRIBE,t) print try: while True: topics,msg = s.recv_multipart() print 'Topic:%s,msg:%s' %(topics,msg) except KeyboardInterrupt: pass print "Done...." if __name__ == "__main__": main() 注意: 这里的发布与订阅角色是绝对的,即发布者无法使用recv,订阅者不能使用send,官网还提供了一种可能出现的问题:当订阅者消费慢于发布, 此时就会出现数据的堆积,而且还是在发布端的堆积(有朋友指出是堆积在消费端,或许是新版本改进,需要读者的尝试和反馈,thx!),显然, 这是不可以被接受的。至于解决方案,或许后面的"分而治之"就是吧 测试: pub端: 发布端 #python zmq-server-pubsub-v02.py tcp://127.0.0.1:10001 sub端:订阅端 #python zmq-server-cs-v01.py tcp://127.0.0.1:10001 sports.football 三,push/pull 分而治之模式. 任务发布端代码 #!/usr/bin/env python # coding:utf8 #author: wangqiankun@lashou-inc.com import zmq import random import time context = zmq.Context() #socket to send messages on sender = context.socket(zmq.PUSH) sender.bind('tcp://*:5557') print 'Press Enter when the workers are ready:' _ = raw_input() print "Sending tasks to workers...." #The first messages is "0" and signals start to batch sender.send('0') #Initialize random mumber generator random.seed() #send 100 tasks total_msec = 0 for task_nbr in range(100): #Random workload from 1 to 100 msecs #print task_nbr, workload = random.randint(1,100) total_msec += workload sender.send(str(workload)) print "Total expected cost:%s msec:%s workload:%s" %(total_msec,task_nbr,workload) work端代码如下: #!/usr/bin/env python # coding:utf8 #author: wangqiankun@lashou-inc.com import sys,time,zmq import commands context = zmq.Context() #socket to receive messages on receiver = context.socket(zmq.PULL) receiver.connect('tcp://127.0.0.1:5557') #Socket to send messages to sender = context.socket(zmq.PUSH) sender.connect("tcp://127.0.0.1:5558") #Process tasks forever while True: s = receiver.recv() #Simple progress indicator for the viewer print s, sys.stdout.write("%s ' ' "%s) sys.stdout.flush() #Do the work time.sleep(int(s)*0.001) #Send results to sink sender.send(s) pull端代码如下: #!/usr/bin/env python # coding:utf8 #author: wangqiankun@lashou-inc.com import sys import time import zmq context = zmq.Context() #Socket to receive messages on receiver = context.socket(zmq.PULL) receiver.bind("tcp://*:5558") #Wait for start of batch s = receiver.recv() #Start our clock now tstart = time.time() #Process 100 confirmations total_msec = 0 for task_nbr in range(100): s = receiver.recv() if task_nbr % 10 == 0: print task_nbr, print s, sys.stdout.write(':') else: print s, #print task_nbr, sys.stdout.write('.') #Calculate and report duration of batch tend = time.time() print "Total elapsed time:%d msec "%((tend-tstart)*1000) 注意点: 这种模式与pub/sub模式一样都是单向的,区别有两点: 1,该模式下在没有消费者的情况下,发布者的信息是不会消耗的(由发布者进程维护) 2,多个消费者消费的是同一列信息,假设A得到了一条信息,则B将不再得到 这种模式主要针对在消费者能力不够的情况下,提供的多消费者并行消费解决方案(也算是之前的pub/sub模式的 那个"堵塞问题"的一个解决策略吧) 其实所谓的分就是pull端去抢push端发出来的任务.谁抢着算谁的. 测试: #python zmq-server-pushpull-v03.py #python zmq-work-pushpull-v03.py #python zmq-client-pushpull-v03.py