一、部署一个三节点集群
下面的链接是最快最简单的一种集群部署方法
3分钟部署一个RabbitMQ集群
上的的例子中,没有映射端口,所以没法从宿主机外部连接容器,下面的yml文件中,暴露了端口。
每个容器应用都映射了宿主机的端口,分别是5602,5612,5622
docker compse文件如下
version: '3'services:stats:image: bitnami/rabbitmqenvironment:- RABBITMQ_NODE_TYPE=stats- RABBITMQ_NODE_NAME=rabbit@stats- RABBITMQ_ERL_COOKIE=s3cr3tc00ki3ports:- '15672:15672'- '5602:5672'volumes:- 'rabbitmqstats_data:/bitnami/rabbitmq/mnesia'queue-disc1:image: bitnami/rabbitmqenvironment:- RABBITMQ_NODE_TYPE=queue-disc- RABBITMQ_NODE_NAME=rabbit@queue-disc1- RABBITMQ_CLUSTER_NODE_NAME=rabbit@stats- RABBITMQ_ERL_COOKIE=s3cr3tc00ki3ports:- '5612:5672'volumes:- 'rabbitmqdisc1_data:/bitnami/rabbitmq/mnesia'queue-ram1:image: bitnami/rabbitmqenvironment:- RABBITMQ_NODE_TYPE=queue-ram- RABBITMQ_NODE_NAME=rabbit@queue-ram1- RABBITMQ_CLUSTER_NODE_NAME=rabbit@stats- RABBITMQ_ERL_COOKIE=s3cr3tc00ki3ports:- '5622:5672'volumes:- 'rabbitmqram1_data:/bitnami/rabbitmq/mnesia'volumes:rabbitmqstats_data:driver: localrabbitmqdisc1_data:driver: localrabbitmqram1_data:driver: local
通过docker-compose up命令,就可以启动三个集群的容器了
[root@localhost mycompose]# docker-compose up
二、配置文件
原来的单节点只配置host和port,现在集群节点,就要配置addresses了,如下所示:
server:port: 8080
spring:application:name: rabbitmq-demo#配置rabbitMq 服务器rabbitmq:
#单节点直接可以写host和port
# host: 192.168.56.201
# port: 5672#集群连接写ip和端口addresses: 192.168.56.202:5602,192.168.56.202:5612,192.168.56.202:5622username: userpassword: bitnami#虚拟hostvirtual-host: virtual01template:mandatory: true #当mandatory设置为true时,交换器无法根据自身的类型和路由键找到一个符合条件的队列,那么RabbitMQ会调用Basic.Return命令将消息返回给生产者。当为false时,则直接丢弃消息publisher-confirm-type: correlated #生产者回调确认机制,由回调来确定消息是否发布成功publisher-returns: true #是否开启生产者returnslistener:simple:acknowledge-mode: manual #手动回复方式,一般建议手动回复,即需要我们自己调用对应的ACK方法prefetch: 10 #每个消费者可拉取的,还未ack的消息数量concurrency: 3 #消费端(每个Listener)的最小线程数max-concurrency: 10 #消费端(每个Listener)的最大线程数
三、代码
生产者
和单节点的发送和消费代码一致,没有变化
@Slf4j
@RestController
@RequestMapping("/rabbit")
public class RabbitSendController implements RabbitTemplate.ConfirmCallback, RabbitTemplate.ReturnCallback {private static final String EXCHANGE_NAME = "my_exchange";private static final String ROUTING_KEY = "my_routing";@Autowiredprivate RabbitTemplate rabbitTemplate;/*** 正常发送并被broker接收* @return*/@RequestMapping("send")public String send() {for (int i = 0; i < 10; i++) {OrderInfo orderInfo = new OrderInfo();orderInfo.setAddress("成都市高新区");orderInfo.setOrderId(String.valueOf(i));orderInfo.setProductName("华为P60:" + i);//设置回调关联的一个idString messageId = UUID.randomUUID().toString();log.info("开始发送消息,当前消息关联id为:{}", messageId);CorrelationData correlationData = new CorrelationData(messageId);MessageProperties messageProperties = new MessageProperties();messageProperties.setDeliveryMode(MessageDeliveryMode.PERSISTENT);Message message = MessageBuilder.withBody(new Gson().toJson(orderInfo).getBytes(StandardCharsets.UTF_8)).andProperties(messageProperties).build();//设置ack回调rabbitTemplate.setConfirmCallback(this);//退回消息的回调rabbitTemplate.setReturnCallback(this);rabbitTemplate.convertAndSend(EXCHANGE_NAME, ROUTING_KEY, message, correlationData);}return "ok";}/*** 设置一个非法的路由键,模拟消息被broker退回的情况,前提是* spring.rabbitmq.template.mandatory=true 当mandatory设置为true时,交换器无法根据自身的类型和路由键找到一个符合条件的队列,那么RabbitMQ会调用Basic.Return命令将消息返回给生产者。当为false时,则直接丢弃消息* <p>* spring.rabbitmq.publisher-returns=true 生产者回调确认机制,由回调来确定消息是否发布成功** @return*/@RequestMapping("send-return")public String sendAndReturn() {OrderInfo orderInfo = new OrderInfo();orderInfo.setAddress("成都市高新区");orderInfo.setOrderId("111");orderInfo.setProductName("小米13");//设置回调关联的一个idString messageId = UUID.randomUUID().toString();log.info("开始发送消息,当前消息关联id为:{}", messageId);CorrelationData correlationData = new CorrelationData(messageId);MessageProperties messageProperties = new MessageProperties();messageProperties.setDeliveryMode(MessageDeliveryMode.PERSISTENT);Message message = MessageBuilder.withBody(new Gson().toJson(orderInfo).getBytes(StandardCharsets.UTF_8)).andProperties(messageProperties).build();//设置ack回调rabbitTemplate.setConfirmCallback(this);//退回消息的回调rabbitTemplate.setReturnCallback(this);//下面这个RoutingKey是没有绑定的,所以发不出去rabbitTemplate.convertAndSend(EXCHANGE_NAME, "error.routing", message, correlationData);return "ok";}@Overridepublic void confirm(CorrelationData correlationData, boolean ack, String cause) {if (correlationData == null) {return;}String messageId = correlationData.getId();if (ack) {log.info("【confirm回调方法】,消息发布成功,messageId={}", messageId);} else {log.info("【confirm回调方法】,消息发布失败,messageId={}", messageId);}}@Overridepublic void returnedMessage(Message message, int replyCode, String replyText, String exchange, String routingKey) {log.info("【returnedMessage回调方法】,消息被退回,message={},replyCode:{},replyText:{},exchange:{},routingKey:{}",new String(message.getBody()), replyCode, replyText, exchange, routingKey);}
}
消费者
@Slf4j
@Component
public class RabbitOrderConsumer {private static final String EXCHANGE_NAME = "my_exchange";private static final String QUEUE_NAME = "my_queue";private static final String ROUTING_KEY = "my_routing";@RabbitListener(bindings = {@QueueBinding(value = @Queue(value = QUEUE_NAME, durable = "true"),exchange = @Exchange(value = EXCHANGE_NAME, type = "topic", durable = "true"), key = ROUTING_KEY)})public void handleMessage(Message message, Channel channel, @Header(AmqpHeaders.DELIVERY_TAG) long tag) throws IOException {//上面这个tag是这么写的么,为什么每次传过来都是1?导致channel被重新创建log.info("接收到消息:{},deliveryTag:{}", new String(message.getBody(), StandardCharsets.UTF_8), tag);channel.basicAck(tag, false);}
}
访问地址:http://localhost:8080/rabbit/send,然后就可以发送消息了,输出日志如下:
开始发送消息,当前消息关联id为:18049efe-a624-4288-a8f0-9c28fd776773
开始发送消息,当前消息关联id为:83d93f90-62f4-41cf-af02-03d496812561
开始发送消息,当前消息关联id为:f83257b2-95b6-408e-a5b9-74d0ec9f30b0
开始发送消息,当前消息关联id为:16a7e471-23ba-408b-9095-6add9ad1e270
开始发送消息,当前消息关联id为:152b0fb0-3a22-452d-93fe-662252c2fd8c
开始发送消息,当前消息关联id为:ade4f703-6075-485f-8e34-ec9b95bf59de
开始发送消息,当前消息关联id为:e4511f82-476a-4f4c-b704-4399baadeaf4
接收到消息:{"orderId":"1","productName":"华为P60:1","address":"成都市高新区"},deliveryTag:1
接收到消息:{"orderId":"0","productName":"华为P60:0","address":"成都市高新区"},deliveryTag:1
开始发送消息,当前消息关联id为:d8cd2dd6-bb9e-4d46-bc42-0d96df70748f
开始发送消息,当前消息关联id为:76950a93-5887-43c1-adef-edc1e29e2fab
开始发送消息,当前消息关联id为:f08a7a68-60da-4c5d-b1b8-c9e4d9453969
【confirm回调方法】,消息发布成功,messageId=18049efe-a624-4288-a8f0-9c28fd776773
【confirm回调方法】,消息发布成功,messageId=83d93f90-62f4-41cf-af02-03d496812561
接收到消息:{"orderId":"3","productName":"华为P60:3","address":"成都市高新区"},deliveryTag:2
接收到消息:{"orderId":"2","productName":"华为P60:2","address":"成都市高新区"},deliveryTag:1
接收到消息:{"orderId":"6","productName":"华为P60:6","address":"成都市高新区"},deliveryTag:3
接收到消息:{"orderId":"5","productName":"华为P60:5","address":"成都市高新区"},deliveryTag:2
接收到消息:{"orderId":"9","productName":"华为P60:9","address":"成都市高新区"},deliveryTag:4
接收到消息:{"orderId":"4","productName":"华为P60:4","address":"成都市高新区"},deliveryTag:2
接收到消息:{"orderId":"7","productName":"华为P60:7","address":"成都市高新区"},deliveryTag:3
接收到消息:{"orderId":"8","productName":"华为P60:8","address":"成都市高新区"},deliveryTag:3
【confirm回调方法】,消息发布成功,messageId=f83257b2-95b6-408e-a5b9-74d0ec9f30b0
【confirm回调方法】,消息发布成功,messageId=16a7e471-23ba-408b-9095-6add9ad1e270
【confirm回调方法】,消息发布成功,messageId=152b0fb0-3a22-452d-93fe-662252c2fd8c
【confirm回调方法】,消息发布成功,messageId=ade4f703-6075-485f-8e34-ec9b95bf59de
【confirm回调方法】,消息发布成功,messageId=e4511f82-476a-4f4c-b704-4399baadeaf4
【confirm回调方法】,消息发布成功,messageId=d8cd2dd6-bb9e-4d46-bc42-0d96df70748f
【confirm回调方法】,消息发布成功,messageId=76950a93-5887-43c1-adef-edc1e29e2fab
【confirm回调方法】,消息发布成功,messageId=f08a7a68-60da-4c5d-b1b8-c9e4d9453969
上述代码仓库:https://gitee.com/syk1234/mqdmo
四、后台管理
登录管理后台页面:http://192.168.56.202:15672/
共有三个节点,两个磁盘节点,一个内存节点。如果你还不清楚什么是磁盘节点,什么是内存节点,可以参考【RabbitMQ 实战】08 集群原理剖析
查看连接情况,发现是连接的是节点rabbit@stats节点
查看队列的情况,队列是在rabbit@stats节点上