RabbitMQ(pika模块)

RabbitMQ

基础

关于MQ:

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,而无需专用连接来链接它们。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。

RabbitMQ安装

1
2
3
4
5
6
7
8
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 install erlang
  
安装RabbitMQ
   $ yum -y install rabbitmq-server

启动/停止:

1
systemctl start/stop rabbitmq

安装python-API:

1
2
3
4
5
6
7
pip install pika
or
easy_install pika
or
源码
  
https://pypi.python.org/pypi/pika


API基础操作


先来看看使用RabbitMQ之前,怎么实现消息队列:利用Queue和Thread,每线程往内存里的队列里put一个数,另一个程序再去内存队列里取数。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
import Queue
import threading
 
 
message = Queue.Queue(10)
 
 
def producer(i):
    while True:
        message.put(i)
 
 
def consumer(i):
    while True:
        msg = message.get()
 
 
for i in range(12):
    t = threading.Thread(target=producer, args=(i,))
    t.start()
 
for i in range(10):
    t = threading.Thread(target=consumer, args=(i,))
    t.start()

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
2
3
4
5
6
7
8
9
10
11
12
13
14
import pika
 
# ######################### 生产者 #########################
 
connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
channel = connection.channel()    #创建通道
 
channel.queue_declare(queue='hello')    #队列名称
 
channel.basic_publish(exchange='',
                      routing_key='hello',  #路由名称
                      body='Hello World!'#发送内容
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
import pika
 
# ########################## 消费者 ##########################
 
connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
channel = connection.channel()
 
channel.queue_declare(queue='hello')    #声明,队列名称,和producer创建的重复没有关系
 
def callback(ch, method, properties, body):
    print(" [x] Received %r" % body)
 
channel.basic_consume(callback,         #获取body后执行回调函数
                      queue='hello',
                      no_ack=True)                #自动应答开启,会给MQ服务器发送一个ack:‘已经收到了’。
 
print(' [*] Waiting for messages. To exit press CTRL+C')
channel.start_consuming()

消费者运行起来后会和RabbitMQ建立长连接,一旦生产者放数据到队列里,消费者就能获取到该值,并进行处理。

1
2
[root@localhost ~]# netstat -ntp |grep beam
tcp6       0      0 192.168.136.8:5672      192.168.136.1:52587     ESTABLISHED 1146/beam


消息安全

1、no-ack = False(自动应答关闭)

如果生产者遇到情况(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
22
import pika
#no-ack
########################### 消费者 ##########################
connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
channel = connection.channel()
 
channel.queue_declare(queue='hello')
 
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)    #主动发送ack
    #打印‘ok’后才告诉MQ,这个消息已经处理完了。
 
channel.basic_consume(callback,
                      queue='hello',
                      no_ack=False)     #自动应答关闭,与channel.basic_ack共同使用
 
print(' [*] Waiting for messages. To exit press CTRL+C')
channel.start_consuming()

2、durable  

make message persistent 使消息持久化

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
import pika
 
#durable
########################## 生产者 #########################
 
connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
channel = connection.channel()
 
channel.queue_declare(queue='hello', durable=True#开启持久化
 
channel.basic_publish(exchange='',
                      routing_key='hello',
                      body='Hello World!',
                      properties=pika.BasicProperties(
                          delivery_mode=2, # make message persistent
                      ))
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
import pika
#durable
########################## 消费者 #########################
 
connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
channel = connection.channel()
 
# make message persistent
channel.queue_declare(queue='hello', 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='hello',
                      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奇偶顺序被消费者拿走,例如:消费者1 去队列中获取 奇数 序列的任务,消费者2去队列中获取 偶数 序列的任务。

channel.basic_qos(prefetch_count=1) 表示谁来谁取,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
import pika
 
connection = pika.BlockingConnection(pika.ConnectionParameters(host='10.211.55.4'))
channel = connection.channel()
 
# make message persistent
channel.queue_declare(queue='hello')
 
 
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)    #增加这行
 
channel.basic_consume(callback,
                      queue='hello',
                      no_ack=False)
 
print(' [*] Waiting for messages. To exit press CTRL+C')
channel.start_consuming()


发布&订阅

与消息队列区别:

消息队列中的数据只要被消费一次便消失。

创建队列的数量:

同一份消息,有多少订阅者,就要创建多少个队列。(RabbitMQ实现发布和订阅时,会为每一个订阅者创建一个队列,而发布者发布消息时,会将消息放置在所有相关队列中。)

语法:

exchange type = fanout        #fanout==>输出到很多

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
# ######################### 发布者 #########################
import pika
import sys
 
connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
channel = connection.channel()
 
channel.exchange_declare(exchange='fanout_name',type='fanout')
 
message = ' '.join(sys.argv[1:]) or "info: Hello World!"
channel.basic_publish(exchange='fanout_name'#自命名exchange
                      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
# ########################## 订阅者1 ##########################
 
import pika
 
connection = pika.BlockingConnection(pika.ConnectionParameters(host='192.168.136.8'))
channel = connection.channel()
 
channel.exchange_declare(exchange='fanout_name',type='fanout')    #创建exchange(if not exist)
 
result = channel.queue_declare(exclusive=True)
queue_name = result.method.queue        #获取队列名称
 
channel.queue_bind(exchange='fanout_name',queue=queue_name)    #通过上面两个值绑定队列
 
print(' [*] Waiting for fanout_name. To exit press CTRL+C')
 
def callback(ch, method, properties, body):
    print(" [x] %r" % body)
 
channel.basic_consume(callback,
                      queue=queue_name,
                      no_ack=True)
 
channel.start_consuming()

创建多个订阅者,能更好的体现它的效果。

运行结果总结:

每个订阅者创建一个exchange队列,名称自定,发布者会把数据发送给所有叫这个名字的队列。因为数据只能被消费一次,所以有多少个订阅者,就有多少个队列。


发送到指定(not 固定)队列

之前事例,发送消息时明确指定某个队列并向其中发送消息,RabbitMQ还支持根据关键字发送

1、按关键字寻找队列发送

exchange type = direct

队列绑定关键字,发送者将数据根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。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
# ######################### 生产者 #########################
#关键字发送
import pika
import sys
 
connection = pika.BlockingConnection(pika.ConnectionParameters(
        host='192.168.136.8'))
channel = connection.channel()
 
channel.exchange_declare(exchange='direct_logs',
                         type='direct')
 
 
message = 'Hello World!'
channel.basic_publish(exchange='direct_logs',
                      routing_key="yes", #"yes","no","db"
                      body=message)
print(" [x] Sent %r" % (message))
connection.close()

模拟两个消费者,一个消费者的队列是("yes","db"),另一个消费者队列("no","db")。如果生产者发送的队列关键字是"yes"or"no",其一匹配;如果生产者发送的队列关键字是"db",则两个消费者都能接收到。

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
########################### 消费者1 ##########################
import pika
import sys
  
connection = pika.BlockingConnection(pika.ConnectionParameters(
        host='192.168.136.8'))
channel = connection.channel()
  
channel.exchange_declare(exchange='direct_logs',
                         type='direct')
  
result = channel.queue_declare(exclusive=True)
queue_name = result.method.queue
 
channel.queue_bind(exchange='direct_logs',
                       queue=queue_name,
                       routing_key='yes')
channel.queue_bind(exchange='direct_logs',
                       queue=queue_name,
                       routing_key='db')
  
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()

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
########################### 消费者2 ##########################
import pika
import sys
  
connection = pika.BlockingConnection(pika.ConnectionParameters(
        host='192.168.136.8'))
channel = connection.channel()
  
channel.exchange_declare(exchange='direct_logs',
                         type='direct')
  
result = channel.queue_declare(exclusive=True)
queue_name = result.method.queue
 
channel.queue_bind(exchange='direct_logs',
                       queue=queue_name,
                       routing_key='no')
channel.queue_bind(exchange='direct_logs',
                       queue=queue_name,
                       routing_key='db')
  
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()

2、模糊匹配

 exchange type = topic

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
发送者路由值              队列中
python.topic.cn          python.*  -- 不匹配
python.topic.cn          python.#  -- 匹配
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
# ######################### 生产者 #########################
#模糊匹配
import pika
import sys
 
connection = pika.BlockingConnection(pika.ConnectionParameters(
        host='192.168.136.8'))
channel = connection.channel()
 
channel.exchange_declare(exchange='topic_logs',
                         type='topic')
 
message = 'Hello World!'
channel.basic_publish(exchange='topic_logs',
                      routing_key="python.topic",
                      body=message)
print(" [x] Sent %r" % (message))
connection.close()

消费者1是‘*’匹配,消费者2是‘#’匹配:

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
########################### 消费者1 ##########################
import pika
import sys
 
connection = pika.BlockingConnection(pika.ConnectionParameters(
        host='192.168.136.8'))
channel = connection.channel()
 
channel.exchange_declare(exchange='topic_logs',
                         type='topic')
 
result = channel.queue_declare(exclusive=True)
queue_name = result.method.queue
 
channel.queue_bind(exchange='topic_logs',
                       queue=queue_name,
                       routing_key='python.*')     #只匹配python.后有一个单词的
 
print(' [*] Waiting for topic_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()
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
########################### 消费者2 ##########################
import pika
import sys
 
connection = pika.BlockingConnection(pika.ConnectionParameters(
        host='192.168.136.8'))
channel = connection.channel()
 
channel.exchange_declare(exchange='topic_logs',
                         type='topic')
 
result = channel.queue_declare(exclusive=True)
queue_name = result.method.queue
 
channel.queue_bind(exchange='topic_logs',
                       queue=queue_name,
                       routing_key='python.#')    #匹配python.后所有单词
 
print(' [*] Waiting for topic_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()

从结果得出结论,如果生产者发送的routing_key是:

  • python.topic.cn    -->    只有消费者2能接收到

  • python.cn            -->    消费者1和消费者2都能接收到

  • python.                -->    消费者1和消费者2都能接收到

  • python                 -->    只有消费者2能接收到


网络搜索的概念:

Topic Exchange – 主题式交换器,通过消息的路由关键字和绑定关键字的模式匹配,将消息路由到被绑定的队列中。

这种路由器类型可以被用来支持经典的发布/订阅消息传输模型——使用主题名字空间作为消息寻址模式,将消息传递给那些部分或者全部匹配主题模式的多个消费者。

主题交换器类型的工作方式如下: 绑定关键字用零个或多个标记构成,每一个标记之间用“.”字符分隔

绑定关键字必须用这种形式明确说明,并支持通配符:“*”匹配一个词组,“#”零个或多个词组。

因此绑定关键字“*.stock.#”匹配路由关键字“usd.stock”和“eur.stock.db”,但是不匹配“stock.nasdaq”


参考来源:http://www.cnblogs.com/wupeiqi/














posted @ 2016-01-21 11:37  大亮头  阅读(3862)  评论(0编辑  收藏  举报