zoukankan      html  css  js  c++  java
  • doc

       本篇博客主要介绍如何通过Python来操作管理RabbitMQ消息队列,大家在工作中可能遇到很多类似RabbitMQ这种消息队列的中间件,如:ZeroMQ、ActiveMQ、MetaMQ等,我们学会了如何操作RabbitMQ的话基本上操作其他的队列都是一通百通。

     一、RabbitMQ安装

        RabbitMQ是一个在AMQP基础上完整的,可复用的企业消息系统,它遵循Mozilla Pulic License开源协议。

    MQ全称为Message Queue,消息队列(MQ)是一种应用程序对应用程序的通信方法。应用程序通过读写出入队列的消息(针对应用程序的数据)来通信,而无需专用链接来链接它们。消息传递指的是程序之间通过在消息中发送数据进行通信,而不是通过直接调用彼此来通信,直接调用通常是用于诸如远程过程调用的技术。排队指的是应用程序通过队列来通信。队列的使用除去了接收和发送应用程序同时执行的要求。

    1,yum安装rabbitmq

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    #安装配置epel源
      rpm -ivh http://dl.fedoraproject.org/pub/epel/6/i386/epel-release-6-8.noarch.rpm
     
    #安装Erlang
      yum -y insatll erlang
     
    #安装RabbitMQ
      yum -y install rabbitmq-server
     
    #注意:
       service rabbitmq-server start/stop

    2,安装API

    1
    2
    3
    4
    5
    #pip安装:
      pip install pika
     
    #源码安装:
      https://pypi.python.org/pypi/pika  #官网地址

        之前我们在介绍线程,进程的时候介绍过python中自带的队列用法,下面我们通过一段代码复习一下:

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    #生产者消费者模型,解耦的意思就是两个程序之间,互相没有关联了,互不影响。
    import queue
    import threading
    import time
    q = queue.Queue(20)      #队列里最多存放20个元素
      
    def productor(arg):            #生成者,创建30个线程来请求吃包子,往队列里添加请求元素
        q.put(str(arg) + '- 包子')
      
    for i in range(30):
        t = threading.Thread(target=productor,args=(i,))
        t.start()
      
    def consumer(arg):       #消费者,接收到队列请求以后开始生产包子,来消费队列里的请求
        while True:
            print(arg,q.get())
            time.sleep(2)
      
    for j in range(3):
        t = threading.Thread(target=consumer,args=(j,))
        t.start()

    二、通过Python来操作RabbitMQ队列

         上面我们已经将环境装备好,下面我们通过Pika模块来对Rabbitmq队列来进行操作,对于RabbitMQ来说,生产和消费不再针对内存里的一个Queue对象,而是某台服务器上的RabbitMQ Server实现的消息队列。

    1,基本用法

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    ####################################生产者#####################################
     
    import pika
     
    connection=pika.BlockingConnection(pika.ConnectionParameters(host='192.168.10.131')) 
    #创建一个链接对象,对象中绑定rabbitmq的IP地址
     
     
    channel=connection.channel()        #创建一个频道
     
    channel.queue_declare(queue='name1'#通过这个频道来创建队列,如果MQ中队列存在忽略,没有则创建
     
    channel.basic_publish(exchange='',
                          routing_key='name1',   #指定队列名称
                          body='Hello World!')   #往该队列中发送一个消息
    print(" [x] Sent 'Hello World!'")
    connection.close()                           #发送完关闭链接
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    #####################################消费者######################################
     
    import pika
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.10.131'))
    #创建一个链接对象,对象中绑定rabbitmq的IP地址
     
    channel = connection.channel()         #创建一个频道
     
    channel.queue_declare(queue='name1')   #通过这个频道来创建队列,如果MQ中队列存在忽略,没有则创建
     
    def callback(ch, method, properties, body):   #callback函数负责接收队列里的消息
        print(" [x] Received %r" % body)
     
    channel.basic_consume(callback,              #从队列里去消息
                          queue='name1',         #指定队列名
                          no_ack=True)
     
    print(' [*] Waiting for messages. To exit press CTRL+C')
    channel.start_consuming()

    acknowledgment 消息不丢失

       上面的例子中如果我们将no-ack=False ,那么当消费者遇到情况(its channel is closed, connection is closed, or TCP connection is lost)挂掉了,那么RabbitMQ会重新将该任务添加到队列中。

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    import pika
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.10.131'))
    channel = connection.channel()
     
    channel.queue_declare(queue='name1')
     
    def callback(ch, method, properties, body):
        print(" [x] Received %r" % body)
        import time
        time.sleep(10)
        print('ok')
        ch.basic_ack(delivery_tag = method.delivery_tag)   #向生成者发送消费完毕的确认信息,然后生产者将此条消息同队列里剔除
     
    channel.basic_consume(callback,
                          queue='name1',                             
                          no_ack=False)                    #如果no_ack=False,当消费者down掉了,RabbitMQ会重新将该任务添加到队列中
     
    print(' [*] Waiting for messages. To exit press CTRL+C')
    channel.start_consuming()

      上例如果消费者中断后如果不超过10秒,重新链接的时候数据还在。当超过10秒之后,消费者往生产者发送了ack,重新链接的时候数据将消失。

    durable消息不丢失

        消费者down掉后我们知道怎么处理了,如果我的RabbitMQ服务down掉了该怎么办呢?

    消息队列是可以做持久化,如果我们在生产消息的时候就指定某条消息需要做持久化,那么RabbitMQ发现有问题时,就会将消息保存到硬盘,持久化下来。

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    ####################################生产者#####################################
    #!/usr/bin/env python
      
    import pika
      
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.10.131'))
      
    channel = connection.channel()
      
    channel.queue_declare(queue='name2', durable=True)    #指定队列持久化
      
    channel.basic_publish(exchange='',
                          routing_key='name2',
                          body='Hello World!',
                          properties=pika.BasicProperties(
                              delivery_mode=2,            #指定消息持久化
                          ))
    print(" [x] Sent 'Hello World!'")
    connection.close()
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    #####################################消费者######################################
    #!/usr/bin/env python
    # -*- coding:utf-8 -*-
    import pika
      
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.10.131'))
      
    channel = connection.channel()
      
    channel.queue_declare(queue='name2', durable=True)
      
      
    def callback(ch, method, properties, body):
        print(" [x] Received %r" % body)
        import time
        time.sleep(10)
        print('ok')
        ch.basic_ack(delivery_tag = method.delivery_tag)
      
    channel.basic_consume(callback,
                          queue='name2',
                          no_ack=False)
      
    print(' [*] Waiting for messages. To exit press CTRL+C')
    channel.start_consuming()

    消息获取顺序

        默认消息队列里的数据是按照顺序被消费者拿走的,例如:消费者1去队列中获取奇数序列任务,消费者2去队列中获取偶数序列的任务,消费者1处理的比较快而消费者2处理的比较慢,那么消费者1就会一直处于繁忙的状态,为了解决这个问题在需要加入下面代码:

    channel.basic_qos(prefetch_count=1)  :表示谁来获取,不再按照奇偶数 排列

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    #!/usr/bin/env python
    # -*- coding:utf-8 -*-
    import pika
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='localhost'))
     
    channel = connection.channel()
     
    channel.queue_declare(queue='name1')
     
     
    def callback(ch, method, properties, body):
        print(" [x] Received %r" % body)
        import time
        time.sleep(10)
        print 'ok'
        ch.basic_ack(delivery_tag = method.delivery_tag)
     
    channel.basic_qos(prefetch_count=1)
     
    channel.basic_consume(callback,
                          queue='name1',
                          no_ack=False)
     
    print(' [*] Waiting for messages. To exit press CTRL+C')
    channel.start_consuming()

    2,发布订阅

        发布订阅和简单的消息队列区别在于,发布订阅会将消息发送给所有的订阅者,而消息队列中的数据被消费一次便消失。所以,RabbitMQ实现发布和订阅时,会为每一个订阅者创建一个队列,二发布者发布消息时,会将消息放置在所有相关队列中。

        在RabbitMQ中,所有生产者提交的消息都有Exchange来接收,然后Exchange按照特定的策略转发到Queue进行存储,RabbitMQ提供了四种Exchange:fanout、direct、topic、header。由于header模式在实际工作中用的比较少,下面主要对前三种进行比较。

    exchange type = fanout :任何发送到Fanout Exchange的消息都会被转发到与该Exchange绑定(Binding)的所有Queue上

      ​为了方便理解,应用了上面这张图,可以清晰的看到相互之间的关系,当我们设置成fanout模式时,如何操作请看下面代码:

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    ####################################发布者#####################################
    import pika
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='localhost'))
    channel = connection.channel()
     
    channel.exchange_declare(exchange='test_fanout',
                             type='fanout')
     
    message = '4456'
    channel.basic_publish(exchange='test_fanout',
                          routing_key='',
                          body=message)
    print(' [x] Sent %r' % message)
    connection.close()
    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    27
    ####################################订阅者#####################################
     
    import pika
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='localhost'))
    channel = connection.channel()
     
    channel.exchange_declare(exchange='test_fanout',        #创建一个exchange
                             type='fanout')                 #任何发送到Fanout Exchange的消息都会被转发到与该Exchange绑定(Binding)的所有Queue上
     
    #随机创建队列
    result = channel.queue_declare(exclusive=True)
    queue_name = result.method.queue
     
    #绑定
    channel.queue_bind(exchange='test_fanout',
                       queue=queue_name)                    #exchange绑定后端队列
     
    print('<------------->')
     
    def callback(ch,method,properties,body):
        print(' [x] %r' % body)
     
    channel.basic_consume(callback,
                          queue=queue_name,
                          no_ack=True)
    channel.start_consuming()

    exchange type = direct:任何发送到Direct Exchange的消息都会被转发到RouteKey中指定的Queue上(关键字发送)

       之前事例,发送消息时明确指定了某个队列并向其中发送消息,RabbitMQ还支持根据关键字发送,即:队列绑定关键字,发送者将数据关键字发送到消息Exchange,Exchange根据关键字判定应该将数据发送至指定队列。

     发布者:

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    #!/usr/bin/env python
    import pika
    import sys
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='localhost'))
     
    channel = connection.channel()
     
    channel.exchange_declare(exchange='direct_test',
                             type='direct')
     
    severity = 'info'         #设置一个key,
    message = '99999'
    channel.basic_publish(exchange='direct_test',
                          routing_key=severity,
                          body=message)
    print(" [x] Sent %r:%r" % (severity, message))
    connection.close()

     ​订阅者1:

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    27
    28
    29
    30
    #!/usr/bin/env python
    import pika
    import sys
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='localhost'))
     
    channel = connection.channel()
     
    channel.exchange_declare(exchange='direct_test',
                             type='direct')
     
    result = channel.queue_declare(exclusive=True)
    queue_name = result.method.queue
     
    severities = ['error','info',]      #绑定队列,并发送关键字error,info
    for severity in severities:
        channel.queue_bind(exchange='direct_test',
                           queue=queue_name,
                           routing_key=severity)
     
    print(' [*] Waiting for logs. To exit press CTRL+C')
     
    def callback(ch, method, properties, body):
        print(" [x] %r:%r" % (method.routing_key, body))
     
    channel.basic_consume(callback,
                          queue=queue_name,
                          no_ack=True)
     
    channel.start_consuming()

    订阅者2:

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    27
    28
    29
    30
    #!/usr/bin/env python
    import pika
    import sys
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='localhost'))
     
    channel = connection.channel()
     
    channel.exchange_declare(exchange='direct_test',
                             type='direct')
     
    result = channel.queue_declare(exclusive=True)
    queue_name = result.method.queue
     
    severities = ['error',]
    for severity in severities:
        channel.queue_bind(exchange='direct_test',
                           queue=queue_name,
                           routing_key=severity)
     
    print(' [*] Waiting for logs. To exit press CTRL+C')
     
    def callback(ch, method, properties, body):
        print(" [x] %r:%r" % (method.routing_key, body))
     
    channel.basic_consume(callback,
                          queue=queue_name,
                          no_ack=True)
     
    channel.start_consuming()

        结论:当我们将发布者的key设置成Error的时候两个队列对可以收到Exchange的消息,当我们将key设置成info后,只有订阅者1可以收到Exchange的消息。

     exchange type = topic:任何发送到Topic Exchange的消息都会被转发到所有关心RouteKey中指定话题的Queue上(模糊匹配)

    在topic类型下,可以让队列绑定几个模糊的关键字,之后发送者将数据发送到exchange,exchange将传入"路由值"和"关键字"进行匹配,匹配成功,则将数据发送到指定队列。

    • # :表示可以匹配0个或多个单词;

    • * :表示只能匹配一个单词。

    1
    2
    3
    4
    5
    #发送路由值        队列中
     
    www.cnblogs.com    www.* --->#无法匹配
     
    www.cnblogs.com    www.# --->#匹配成功

    发布者:

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    #!/usr/bin/env python
    import pika
    import sys
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='localhost'))
     
    channel = connection.channel()
     
    channel.exchange_declare(exchange='topic_logs',
                             type='topic')
     
    routing_key = sys.argv[1] if len(sys.argv) > 1 else 'anonymous.info'
     
    message = ' '.join(sys.argv[2:]) or 'Hello World!'
     
    channel.basic_publish(exchange='topic_logs',
                          routing_key=routing_key,
                          body=message)
    print(" [x] Sent %r:%r" % (routing_key, message))
     
    connection.close()
     
     
    #执行方式:
    python xxx.py name1   #name1为routing_key

    订阅者:

    1
    2
    3
    4
    5
    6
    7
    8
    9
    10
    11
    12
    13
    14
    15
    16
    17
    18
    19
    20
    21
    22
    23
    24
    25
    26
    27
    28
    29
    30
    31
    32
    33
    34
    35
    36
    37
    #!/usr/bin/env python
    import pika
    import sys
     
    connection = pika.BlockingConnection(pika.ConnectionParameters(host='localhost'))
     
    channel = connection.channel()
     
    channel.exchange_declare(exchange='topic_logs',
                             type='topic')
     
    result = channel.queue_declare(exclusive=True)
    queue_name = result.method.queue
     
    binding_keys = sys.argv[1:]
    if not binding_keys:
        sys.stderr.write("Usage: %s [binding_key]... " % sys.argv[0])
        sys.exit(1)
     
    for binding_key in binding_keys:
        channel.queue_bind(exchange='topic_logs',
                           queue=queue_name,
                           routing_key=binding_key)
     
    print(' [*] Waiting for logs. To exit press CTRL+C')
     
    def callback(ch, method, properties, body):
        print(" [x] %r:%r" % (method.routing_key, body))
     
    channel.basic_consume(callback,
                          queue=queue_name,
                          no_ack=True)
     
    channel.start_consuming()
     
    #执行方式:
    python xxx,py name1

    更多相关内容请参考以下连接:

    http://www.rabbitmq.com/documentation.html

    http://blog.csdn.net/songfreeman/article/details/50945025

  • 相关阅读:
    HDU4474 Yet Another Multiple Problem BFS搜索
    HDU4473 Exam 数学分析
    2013ACM多校联合(4)
    POJ1273 网络流...
    HDU4472 Count 递推
    POJ1149 PIGS 网络流
    UVA10881 Piotr's Ants 想法题
    javascript js string.Format()收集
    修改 设置 vs.net 网站 调试 设为 起始页
    【转】HTML5杂谈 概念与现行游戏 割绳子 宝石迷阵
  • 原文地址:https://www.cnblogs.com/phennry/p/5713274.html
Copyright © 2011-2022 走看看