rabbitmq-----发布订阅模式
模型组成
一个消费者Producer,一个交换机Exchange,多个消息队列Queue,多个消费者Consumer
一个生产者,多个消费者,每一个消费者都有自己的一个队列,生产者没有将消息直接发送到队列,而是发送到了交换机,每个队列绑定交换机,生产者发送的消息经过交换机,到达队列,实现一个消息被多个消费者获取的目的。需要注意的是,如果将消息发送到一个没有队列绑定的exchange上面,那么该消息将会丢失,这是因为在rabbitMQ中exchange不具备存储消息的能力,只有队列具备存储消息的能力。
Exchange
相比较于前两种模型Hello World和Work,这里多一个一个Exchange。其实Exchange是RabbitMQ的标配组成部件之一,前两种没有提到Exchange是为了简化模型,即使模型中没有看到Exchange的声明,其实还是声明了一个默认的Exchange。
RabbitMQ中实际发送消息并不是直接将消息发送给消息队列,消息队列也没那么聪明知道这条消息从哪来要到哪去。RabbitMQ会先将消息发送个Exchange,Exchange会根据这条消息打上的标记知道该条消息从哪来到哪去。
Exchange凭什么知道消息的何去何从,因为Exchange有几种类型:direct,fanout,topic和headers。这里说的订阅者模式就可以认为是fanout模式了。
RabbitMQ
中,所有生产者提交的消息都由Exchange
来接受,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 .发布/订阅模式就是是基于
fanout Exchange
实现的。
fanout
这种模式不需要指定队列名称,需要将Exchange
和queue
绑定,他们之间的关系是‘多对多’的关系
任何发送到fanout Exchange
的消息都会被转发到与该Exchange
绑定的queue
上面。
订阅者模式有何不同
订阅者模式相对前面的Work模式有和不同?Work也有多个消费者,但是只有一个消息队列,并且一个消息只会被某一个消费者消费。但是订阅者模式不一样,它有多个消息队列,也有多个消费者,而且一条消息可以被多个消费者消费,类似广播模式。下面通过实例代码看看这种模式是如何收发消息的。
1 package com.maozw.mq.pubsub; 2 3 import com.maozw.mq.config.RabbitConfig; 4 import com.rabbitmq.client.Channel; 5 import org.slf4j.Logger; 6 import org.slf4j.LoggerFactory; 7 import org.springframework.amqp.rabbit.connection.Connection; 8 import org.springframework.amqp.rabbit.connection.ConnectionFactory; 9 import org.springframework.beans.factory.annotation.Autowired; 10 import org.springframework.web.bind.annotation.PathVariable; 11 import org.springframework.web.bind.annotation.RequestMapping; 12 import org.springframework.web.bind.annotation.RestController; 13 14 import java.io.IOException; 15 import java.util.concurrent.TimeoutException; 16 17 import static org.apache.log4j.varia.ExternallyRolledFileAppender.OK; 18 19 /** 20 * work 模式 21 * 两种分发: 轮询分发 + 公平分发 22 * 轮询分发:消费端:自动确认消息;boolean autoAck = true; 23 * 公平分发: 消费端:手动确认消息 boolean autoAck = false; channel.basicAck(envelope.getDeliveryTag(),false); 24 * 25 * @author MAOZW 26 * @Description: ${todo} 27 * @date 2018/11/26 15:06 28 */ 29 @RestController 30 @RequestMapping("/publish") 31 public class PublishProducer { 32 private static final Logger LOGGER = LoggerFactory.getLogger(PublishProducer.class); 33 @Autowired 34 RabbitConfig rabbitConfig; 35 36 37 @RequestMapping("/send/{exchangeName}/{queueName}") 38 public String send(@PathVariable String exchangeName, @PathVariable String queueName) throws IOException, TimeoutException { 39 Connection connection = null; 40 Channel channel= null; 41 try { 42 ConnectionFactory connectionFactory = rabbitConfig.connectionFactory(); 43 connection = connectionFactory.createConnection(); 44 channel = connection.createChannel(false); 45 46 /** 47 * 申明交换机 48 */ 49 channel.exchangeDeclare(exchangeName,"fanout"); 50 51 /** 52 * 发送消息 53 * 每个消费者 发送确认消息之前,消息队列不会发送下一个消息给消费者,一次只处理一个消息 54 * 自动模式无需设置下面设置 55 */ 56 int prefetchCount = 1; 57 channel.basicQos(prefetchCount); 58 59 String Hello = ">>>> Hello Simple <<<<"; 60 for (int i = 0; i < 5; i++) { 61 String message = Hello + i; 62 channel.basicPublish(RabbitConfig.EXCHANGE_AAAAA, "", null, message.getBytes()); 63 LOGGER.info("生产消息: " + message); 64 } 65 return "OK"; 66 }catch (Exception e) { 67 68 } finally { 69 connection.close(); 70 channel.close(); 71 return OK; 72 } 73 } 74 }
订阅1
1 package com.maozw.mq.pubsub; 2 3 import com.maozw.mq.config.RabbitConfig; 4 import com.rabbitmq.client.AMQP; 5 import com.rabbitmq.client.Channel; 6 import com.rabbitmq.client.DefaultConsumer; 7 import com.rabbitmq.client.Envelope; 8 import org.slf4j.Logger; 9 import org.slf4j.LoggerFactory; 10 import org.springframework.amqp.rabbit.connection.Connection; 11 import org.springframework.amqp.rabbit.connection.ConnectionFactory; 12 13 import java.io.IOException; 14 15 /** 16 * @author MAOZW 17 * @Description: ${todo} 18 * @date 2018/11/26 15:06 19 */ 20 21 public class SubscribeConsumer { 22 private static final Logger LOGGER = LoggerFactory.getLogger(SubscribeConsumer.class); 23 24 public static void main(String[] args) throws IOException { 25 ConnectionFactory connectionFactory = RabbitConfig.getConnectionFactory(); 26 Connection connection = connectionFactory.createConnection(); 27 Channel channel = connection.createChannel(false); 28 /** 29 * 创建队列申明 30 */ 31 boolean durable = true; 32 channel.queueDeclare(RabbitConfig.QUEUE_PUBSUB_FANOUT, durable, false, false, null); 33 /** 34 * 绑定队列到交换机 35 */ 36 channel.queueBind(RabbitConfig.QUEUE_PUBSUB_FANOUT, RabbitConfig.EXCHANGE_AAAAA,""); 37 38 /** 39 * 改变分发规则 40 */ 41 channel.basicQos(1); 42 DefaultConsumer consumer = new DefaultConsumer(channel) { 43 @Override 44 public void handleDelivery(String consumerTag, Envelope envelope, AMQP.BasicProperties properties, byte[] body) throws IOException { 45 super.handleDelivery(consumerTag, envelope, properties, body); 46 System.out.println("[2] 接口数据 : " + new String(body, "utf-8")); 47 try { 48 Thread.sleep(300); 49 } catch (InterruptedException e) { 50 e.printStackTrace(); 51 } finally { 52 System.out.println("[2] done!"); 53 //消息应答:手动回执,手动确认消息 54 channel.basicAck(envelope.getDeliveryTag(),false); 55 } 56 } 57 }; 58 //监听队列 59 /** 60 * autoAck 消息应答 61 * 默认轮询分发打开:true :这种模式一旦rabbitmq将消息发送给消费者,就会从内存中删除该消息,不关心客户端是否消费正常。 62 * 使用公平分发需要关闭autoAck:false 需要手动发送回执 63 */ 64 boolean autoAck = false; 65 channel.basicConsume(RabbitConfig.QUEUE_PUBSUB_FANOUT,autoAck, consumer); 66 } 67 68 }
1 package com.maozw.mq.pubsub; 2 3 import com.maozw.mq.config.RabbitConfig; 4 import com.rabbitmq.client.AMQP; 5 import com.rabbitmq.client.Channel; 6 import com.rabbitmq.client.DefaultConsumer; 7 import com.rabbitmq.client.Envelope; 8 import org.slf4j.Logger; 9 import org.slf4j.LoggerFactory; 10 import org.springframework.amqp.rabbit.connection.Connection; 11 import org.springframework.amqp.rabbit.connection.ConnectionFactory; 12 13 import java.io.IOException; 14 15 /** 16 * @author MAOZW 17 * @Description: ${todo} 18 * @date 2018/11/26 15:06 19 */ 20 21 public class SubscribeConsumer2 { 22 private static final Logger LOGGER = LoggerFactory.getLogger(SubscribeConsumer2.class); 23 24 public static void main(String[] args) throws IOException { 25 ConnectionFactory connectionFactory = RabbitConfig.getConnectionFactory(); 26 Connection connection = connectionFactory.createConnection(); 27 Channel channel = connection.createChannel(false); 28 /** 29 * 创建队列申明 30 */ 31 boolean durable = true; 32 channel.queueDeclare(RabbitConfig.QUEUE_PUBSUB_FANOUT2, durable, false, false, null); 33 /** 34 * 绑定队列到交换机 35 */ 36 channel.queueBind(RabbitConfig.QUEUE_PUBSUB_FANOUT2, RabbitConfig.EXCHANGE_AAAAA,""); 37 38 /** 39 * 改变分发规则 40 */ 41 channel.basicQos(1); 42 DefaultConsumer consumer = new DefaultConsumer(channel) { 43 @Override 44 public void handleDelivery(String consumerTag, Envelope envelope, AMQP.BasicProperties properties, byte[] body) throws IOException { 45 super.handleDelivery(consumerTag, envelope, properties, body); 46 System.out.println("[2] 接口数据 : " + new String(body, "utf-8")); 47 try { 48 Thread.sleep(400); 49 } catch (InterruptedException e) { 50 e.printStackTrace(); 51 } finally { 52 System.out.println("[2] done!"); 53 //消息应答:手动回执,手动确认消息 54 channel.basicAck(envelope.getDeliveryTag(),false); 55 } 56 } 57 }; 58 //监听队列 59 /** 60 * autoAck 消息应答 61 * 默认轮询分发打开:true :这种模式一旦rabbitmq将消息发送给消费者,就会从内存中删除该消息,不关心客户端是否消费正常。 62 * 使用公平分发需要关闭autoAck:false 需要手动发送回执 63 */ 64 boolean autoAck = false; 65 channel.basicConsume(RabbitConfig.QUEUE_PUBSUB_FANOUT2,autoAck, consumer); 66 } 67 }