RabbitMQ 七种工作模式介绍（下） 
路由模式的升级版, 在routingKey的基础上,增加了通配符的功能, 使之更加灵活.
Topics和Routing的基本原理相同，即：⽣产者将消息发给交换机，交换机根据RoutingKey将消息转发给与RoutingKey匹配的队列. 类似于正则表达式的⽅式来定义Routingkey的模式.
适合场景: 需要灵活匹配和过滤消息的场景
生产者：
243
0public static void main(String[] args) throws IOException, TimeoutException {
//1. 建立连接
ConnectionFactory connectionFactory = new ConnectionFactory();
connectionFactory.setHost(Constants.HOST);
connectionFactory.setPort(Constants.PORT); //需要提前开放端口号
connectionFactory.setUsername(Constants.USER_NAME);//账号
connectionFactory.setPassword(Constants.PASSWORD); //密码
connectionFactory.setVirtualHost(Constants.VIRTUAL_HOST); //虚拟主机
Connection connection = connectionFactory.newConnection();
//2. 开启信道
Channel channel = connection.createChannel();
//3. 声明交换机
channel.exchangeDeclare(Constants.TOPIC_EXCHANGE, BuiltinExchangeType.TOPIC, true);
//4. 声明队列
channel.queueDeclare(Constants.TOPIC_QUEUE1,true,false,false,null);
channel.queueDeclare(Constants.TOPIC_QUEUE2,true,false,false,null);
//5. 交换机和队列绑定
channel.queueBind(Constants.TOPIC_QUEUE1,Constants.TOPIC_EXCHANGE,"*.a.*");
channel.queueBind(Constants.TOPIC_QUEUE2,Constants.TOPIC_EXCHANGE,"*.*.b");
channel.queueBind(Constants.TOPIC_QUEUE2,Constants.TOPIC_EXCHANGE,"c.#");
//6. 发布消息
String msg_a = "hello 路由队列~ my routingKey is ae.a.f...";
channel.basicPublish(Constants.TOPIC_EXCHANGE,"ae.a.f", null, msg_a.getBytes());
String msg_b = "hello 路由队列~ my routingKey is ef.a.b...";
channel.basicPublish(Constants.TOPIC_EXCHANGE,"ef.a.b", null, msg_b.getBytes());
String msg_c = "hello 路由队列~ my routingKey is c.ef.d ...";
channel.basicPublish(Constants.TOPIC_EXCHANGE,"c.ef.d", null, msg_c.getBytes());
System.out.println("消息发送成功");
//7. 释放资源
channel.close();
connection.close();
}
消费者1： 只接收*.a.*
消费者2： 接收*.*.b c.#
在RPC通信的过程中, 没有⽣产者和消费者, ⽐较像咱们RPC远程调⽤, ⼤概就是通过两个队列实现了⼀个可回调的过程.
客户端：1.发送请求（携带replyTo,correlationId ）2.接收响应（验证correlationId）
服务器：1.接收请求 进行响应 2.发送响应（按客户端指定的replyTo，设置correlationId）
客户端：
public static void main(String[] args) throws IOException, TimeoutException, InterruptedException {
//1. 建立连接
ConnectionFactory connectionFactory = new ConnectionFactory();
connectionFactory.setHost(Constants.HOST);
connectionFactory.setPort(Constants.PORT); //需要提前开放端口号
connectionFactory.setUsername(Constants.USER_NAME);//账号
connectionFactory.setPassword(Constants.PASSWORD); //密码
connectionFactory.setVirtualHost(Constants.VIRTUAL_HOST); //虚拟主机
Connection connection = connectionFactory.newConnection();
//2. 开启信道
Channel channel = connection.createChannel();
channel.queueDeclare(Constants.RPC_RESPONSE_QUEUE,true,false,false,null);
channel.queueDeclare(Constants.RPC_REQUEST_QUEUE,true,false,false,null);
//3. 发送请求
String msg = "hello rpc...";
//设置请求的唯一标识
String correlationID = UUID.randomUUID().toString();
//设置请求的相关属性
AMQP.BasicProperties props = new AMQP.BasicProperties().builder()
.correlationId(correlationID)
.replyTo(Constants.RPC_RESPONSE_QUEUE)
.build();
channel.basicPublish("",Constants.RPC_REQUEST_QUEUE,props,msg.getBytes());
//4. 接收响应
//使用阻塞队列, 来存储响应信息
final BlockingQueue<String> response = new ArrayBlockingQueue<>(1);
DefaultConsumer consumer = new DefaultConsumer(channel){
@Override
public void handleDelivery(String consumerTag, Envelope envelope, AMQP.BasicProperties properties, byte[] body) throws IOException {
String respMsg = new String(body);
System.out.println("接收到回调消息: "+ respMsg);
if (correlationID.equals(properties.getCorrelationId())){
//如果correlationID校验一致
response.offer(respMsg);
}
}
};
channel.basicConsume(Constants.RPC_RESPONSE_QUEUE, true, consumer);
String result = response.take();
System.out.println("[RPC Client 响应结果]:"+ result);
}
客户端：
public static void main(String[] args) throws IOException, TimeoutException, InterruptedException {
//1. 建立连接
ConnectionFactory connectionFactory = new ConnectionFactory();
connectionFactory.setHost(Constants.HOST);
connectionFactory.setPort(Constants.PORT); //需要提前开放端口号
connectionFactory.setUsername(Constants.USER_NAME);//账号
connectionFactory.setPassword(Constants.PASSWORD); //密码
connectionFactory.setVirtualHost(Constants.VIRTUAL_HOST); //虚拟主机
Connection connection = connectionFactory.newConnection();
//2. 开启信道
Channel channel = connection.createChannel();
channel.queueDeclare(Constants.RPC_RESPONSE_QUEUE,true,false,false,null);
channel.queueDeclare(Constants.RPC_REQUEST_QUEUE,true,false,false,null);
//3. 发送请求
String msg = "hello rpc...";
//设置请求的唯一标识
String correlationID = UUID.randomUUID().toString();
//设置请求的相关属性
AMQP.BasicProperties props = new AMQP.BasicProperties().builder()
.correlationId(correlationID)
.replyTo(Constants.RPC_RESPONSE_QUEUE)
.build();
channel.basicPublish("",Constants.RPC_REQUEST_QUEUE,props,msg.getBytes());
//4. 接收响应
//使用阻塞队列, 来存储响应信息
final BlockingQueue<String> response = new ArrayBlockingQueue<>(1);
DefaultConsumer consumer = new DefaultConsumer(channel){
@Override
public void handleDelivery(String consumerTag, Envelope envelope, AMQP.BasicProperties properties, byte[] body) throws IOException {
String respMsg = new String(body);
System.out.println("接收到回调消息: "+ respMsg);
if (correlationID.equals(properties.getCorrelationId())){
//如果correlationID校验一致
response.offer(respMsg);
}
}
};
channel.basicConsume(Constants.RPC_RESPONSE_QUEUE, true, consumer);
String result = response.take();
System.out.println("[RPC Client 响应结果]:"+ result);
}
消息丢失其中一种情况 ⽣产者问题. 因为应⽤程序故障, ⽹络抖动等各种原因, ⽣产者没有成功向broker发送消息
可以采⽤发布确认(Publisher Confirms)机制实现
发送⽅确认机制最⼤的好处在于它是异步的, ⽣产者可以同时发布消息和等待信道返回确认消息.
1. 当消息最终得到确认之后, ⽣产者可以通过回调⽅法来处理该确认消息.
2. 如果RabbitMQ因为⾃⾝内部错误导致消息丢失, 就会发送⼀条nack(Basic.Nack)命令, ⽣产者同样 可以在回调⽅法中处理该nack命令.
使⽤发送确认机制, 必须要信道设置成confirm(确认)模式.
发布确认有3种策略, 接下来我们来学习这三种策略
private static void publishingMessagesIndividually() throws Exception {
try(Connection connection = createConnection()) {
//1.开启信道
Channel channel = connection.createChannel();
//2. 设置信道为confirm模式
channel.confirmSelect();
//3.声明队列
channel.queueDeclare(Constants.PUBLISHER_CONFIRMS_QUEUE1,true,false,false,null);
//4.发送消息，并等待确认
Long start = System.currentTimeMillis();
for (int i = 0; i < MESSAGE_COUNT; i++) {
String msg = "hello publisher confirms"+i;
channel.basicPublish("",Constants.PUBLISHER_CONFIRMS_QUEUE1,null,msg.getBytes());
//等待确认
channel.waitForConfirmsOrDie(5000);
}
Long end = System.currentTimeMillis();
System.out.printf("单独确认策略, 消息条数: %d, 耗时: %d ms \n",MESSAGE_COUNT, end-start);
}
}
private static void publishingMessagesInBatches() throws Exception {
try(Connection connection = createConnection()) {
//1.开启信道
Channel channel = connection.createChannel();
//2. 设置信道为confirm模式
channel.confirmSelect();
//3.声明队列
channel.queueDeclare(Constants.PUBLISHER_CONFIRMS_QUEUE2,true,false,false,null);
//4. 发送消息, 并进行确认
long start = System.currentTimeMillis();
int batchSize = 100;
int outstandingMessageCount = 0;
for (int i = 0; i < MESSAGE_COUNT; i++) {
String msg = "hello publisher confirms"+i;
channel.basicPublish("",Constants.PUBLISHER_CONFIRMS_QUEUE2,null,msg.getBytes());
outstandingMessageCount++;
if(outstandingMessageCount == batchSize) {
channel.waitForConfirmsOrDie(5000);
outstandingMessageCount = 0;
}
}
if(outstandingMessageCount > 0) {
channel.waitForConfirmsOrDie(5000);
}
long end = System.currentTimeMillis();
System.out.printf("批量确认策略, 消息条数: %d, 耗时: %d ms \n",MESSAGE_COUNT, end-start);
}
}
private static void handlingPublisherConfirmsAsynchronously() throws Exception {
try(Connection connection = createConnection()) {
//1.开启信道
Channel channel = connection.createChannel();
//2. 设置信道为confirm模式
channel.confirmSelect();
//3.声明队列
channel.queueDeclare(Constants.PUBLISHER_CONFIRMS_QUEUE3,true,false,false,null);
//4. 监听confirm
//集合中存储的是未确认的消息ID
long start = System.currentTimeMillis();
SortedSet<Long> confirmSeqNo = Collections.synchronizedSortedSet(new TreeSet<>());
channel.addConfirmListener(new ConfirmListener() {
@Override
public void handleAck(long deliveryTag, boolean multiple) throws IOException {
if(multiple) {
confirmSeqNo.headSet(deliveryTag+1).clear();
} else {
confirmSeqNo.remove(deliveryTag);
}
}
@Override
public void handleNack(long deliveryTag, boolean multiple) throws IOException {
if(multiple) {
confirmSeqNo.headSet(deliveryTag+1).clear();
}else {
confirmSeqNo.remove(deliveryTag);
}
//业务需要根据实际场景进行处理, 比如重发, 此处代码省略
}
});
for (int i = 0; i < MESSAGE_COUNT; i++) {
String msg = "hello publisher confirms"+i;
long seqNo = channel.getNextPublishSeqNo();
channel.basicPublish("",Constants.PUBLISHER_CONFIRMS_QUEUE3,null,msg.getBytes());
confirmSeqNo.add(seqNo);
}
while (!confirmSeqNo.isEmpty()) {
Thread.sleep(10);
}
long end = System.currentTimeMillis();
System.out.printf("异步确认策略, 消息条数: %d, 耗时: %d ms \n",MESSAGE_COUNT, end-start);
}
}
消息数越多, 异步确认的优势越明显
@Copyrights 2016-2023 杭州玳数科技有限公司 浙ICP备15044486号-1 浙公网安备33011002011932号
