RocketMQ事务消费和顺序消费详解
一、RocketMq有3中消息类型
1.普通消费
2. 顺序消费
3.事务消费
- 顺序消费场景
在网购的时候,我们需要下单,那么下单需要假如有三个顺序,第一、创建订单 ,第二:订单付款,第三:订单完成。也就是这个三个环节要有顺序,这个订单才有意义。RocketMQ可以保证顺序消费。
- rocketMq实现顺序消费的原理
produce在发送消息的时候,把消息发到同一个队列(queue)中,消费者注册消息监听器为MessageListenerOrderly,这样就可以保证消费端只有一个线程去消费消息
注意:是把把消息发到同一个队列(queue),不是同一个topic,默认情况下一个topic包括4个queue
单个节点(Producer端1个、Consumer端1个)
1、Producer.java
package order; import java.util.List; import com.alibaba.rocketmq.client.exception.MQBrokerException; import com.alibaba.rocketmq.client.exception.MQClientException; import com.alibaba.rocketmq.client.producer.DefaultMQProducer; import com.alibaba.rocketmq.client.producer.MessageQueueSelector; import com.alibaba.rocketmq.client.producer.SendResult; import com.alibaba.rocketmq.common.message.Message; import com.alibaba.rocketmq.common.message.MessageQueue; import com.alibaba.rocketmq.remoting.exception.RemotingException; /** * Producer,发送顺序消息 */ public class Producer { public static void main(String[] args) { try { DefaultMQProducer producer = new DefaultMQProducer("order_Producer"); producer.setNamesrvAddr("192.168.100.145:9876;192.168.100.146:9876;192.168.100.149:9876;192.168.100.239:9876"); producer.start(); // String[] tags = new String[] { "TagA", "TagB", "TagC", "TagD", // "TagE" }; for (int i = 1; i <= 5; i++) { Message msg = new Message("TopicOrderTest", "order_1", "KEY" + i, ("order_1 " + i).getBytes()); SendResult sendResult = producer.send(msg, new MessageQueueSelector() { public MessageQueue select(List<MessageQueue> mqs, Message msg, Object arg) { Integer id = (Integer) arg; int index = id % mqs.size(); return mqs.get(index); } }, 0); System.out.println(sendResult); } producer.shutdown(); } catch (MQClientException e) { e.printStackTrace(); } catch (RemotingException e) { e.printStackTrace(); } catch (MQBrokerException e) { e.printStackTrace(); } catch (InterruptedException e) { e.printStackTrace(); } } }
2、Consumer.java
package order; import java.util.List; import java.util.concurrent.TimeUnit; import java.util.concurrent.atomic.AtomicLong; import com.alibaba.rocketmq.client.consumer.DefaultMQPushConsumer; import com.alibaba.rocketmq.client.consumer.listener.ConsumeOrderlyContext; import com.alibaba.rocketmq.client.consumer.listener.ConsumeOrderlyStatus; import com.alibaba.rocketmq.client.consumer.listener.MessageListenerOrderly; import com.alibaba.rocketmq.client.exception.MQClientException; import com.alibaba.rocketmq.common.consumer.ConsumeFromWhere; import com.alibaba.rocketmq.common.message.MessageExt; /** * 顺序消息消费,带事务方式(应用可控制Offset什么时候提交) */ public class Consumer1 { public static void main(String[] args) throws MQClientException { DefaultMQPushConsumer consumer = new DefaultMQPushConsumer("order_Consumer"); consumer.setNamesrvAddr("192.168.100.145:9876;192.168.100.146:9876;192.168.100.149:9876;192.168.100.239:9876"); /** * 设置Consumer第一次启动是从队列头部开始消费还是队列尾部开始消费<br> * 如果非第一次启动,那么按照上次消费的位置继续消费 */ consumer.setConsumeFromWhere(ConsumeFromWhere.CONSUME_FROM_FIRST_OFFSET); consumer.subscribe("TopicOrderTest", "*"); consumer.registerMessageListener(new MessageListenerOrderly() { AtomicLong consumeTimes = new AtomicLong(0); public ConsumeOrderlyStatus consumeMessage(List<MessageExt> msgs, ConsumeOrderlyContext context) { // 设置自动提交 context.setAutoCommit(true); for (MessageExt msg : msgs) { System.out.println(msg + ",内容:" + new String(msg.getBody())); } try { TimeUnit.SECONDS.sleep(5L); } catch (InterruptedException e) { e.printStackTrace(); } ; return ConsumeOrderlyStatus.SUCCESS; } }); consumer.start(); System.out.println("Consumer1 Started."); } }
结果如下图所示:
这个五条数据被顺序消费了
- 多个节点(Producer端1个、Consumer端2个)
Producer.java
package order; import java.util.List; import com.alibaba.rocketmq.client.exception.MQBrokerException; import com.alibaba.rocketmq.client.exception.MQClientException; import com.alibaba.rocketmq.client.producer.DefaultMQProducer; import com.alibaba.rocketmq.client.producer.MessageQueueSelector; import com.alibaba.rocketmq.client.producer.SendResult; import com.alibaba.rocketmq.common.message.Message; import com.alibaba.rocketmq.common.message.MessageQueue; import com.alibaba.rocketmq.remoting.exception.RemotingException; /** * Producer,发送顺序消息 */ public class Producer { public static void main(String[] args) { try { DefaultMQProducer producer = new DefaultMQProducer("order_Producer"); producer.setNamesrvAddr("192.168.100.145:9876;192.168.100.146:9876;192.168.100.149:9876;192.168.100.239:9876"); producer.start(); // String[] tags = new String[] { "TagA", "TagB", "TagC", "TagD", // "TagE" }; for (int i = 1; i <= 5; i++) { Message msg = new Message("TopicOrderTest", "order_1", "KEY" + i, ("order_1 " + i).getBytes()); SendResult sendResult = producer.send(msg, new MessageQueueSelector() { public MessageQueue select(List<MessageQueue> mqs, Message msg, Object arg) { Integer id = (Integer) arg; int index = id % mqs.size(); return mqs.get(index); } }, 0); System.out.println(sendResult); } for (int i = 1; i <= 5; i++) { Message msg = new Message("TopicOrderTest", "order_2", "KEY" + i, ("order_2 " + i).getBytes()); SendResult sendResult = producer.send(msg, new MessageQueueSelector() { public MessageQueue select(List<MessageQueue> mqs, Message msg, Object arg) { Integer id = (Integer) arg; int index = id % mqs.size(); return mqs.get(index); } }, 1); System.out.println(sendResult); } for (int i = 1; i <= 5; i++) { Message msg = new Message("TopicOrderTest", "order_3", "KEY" + i, ("order_3 " + i).getBytes()); SendResult sendResult = producer.send(msg, new MessageQueueSelector() { public MessageQueue select(List<MessageQueue> mqs, Message msg, Object arg) { Integer id = (Integer) arg; int index = id % mqs.size(); return mqs.get(index); } }, 2); System.out.println(sendResult); } producer.shutdown(); } catch (MQClientException e) { e.printStackTrace(); } catch (RemotingException e) { e.printStackTrace(); } catch (MQBrokerException e) { e.printStackTrace(); } catch (InterruptedException e) { e.printStackTrace(); } } }
Consumer1.java
/** * 顺序消息消费,带事务方式(应用可控制Offset什么时候提交) */ public class Consumer1 { public static void main(String[] args) throws MQClientException { DefaultMQPushConsumer consumer = new DefaultMQPushConsumer("order_Consumer"); consumer.setNamesrvAddr("192.168.100.145:9876;192.168.100.146:9876;192.168.100.149:9876;192.168.100.239:9876"); /** * 设置Consumer第一次启动是从队列头部开始消费还是队列尾部开始消费<br> * 如果非第一次启动,那么按照上次消费的位置继续消费 */ consumer.setConsumeFromWhere(ConsumeFromWhere.CONSUME_FROM_FIRST_OFFSET); consumer.subscribe("TopicOrderTest", "*"); /** * 实现了MessageListenerOrderly表示一个队列只会被一个线程取到 *,第二个线程无法访问这个队列 */ consumer.registerMessageListener(new MessageListenerOrderly() { AtomicLong consumeTimes = new AtomicLong(0); public ConsumeOrderlyStatus consumeMessage(List<MessageExt> msgs, ConsumeOrderlyContext context) { // 设置自动提交 context.setAutoCommit(true); for (MessageExt msg : msgs) { System.out.println(msg + ",内容:" + new String(msg.getBody())); } try { TimeUnit.SECONDS.sleep(5L); } catch (InterruptedException e) { e.printStackTrace(); } ; return ConsumeOrderlyStatus.SUCCESS; } }); consumer.start(); System.out.println("Consumer1 Started."); } }
Consumer2.java
/** * 顺序消息消费,带事务方式(应用可控制Offset什么时候提交) */ public class Consumer2 { public static void main(String[] args) throws MQClientException { DefaultMQPushConsumer consumer = new DefaultMQPushConsumer("order_Consumer"); consumer.setNamesrvAddr("192.168.100.145:9876;192.168.100.146:9876;192.168.100.149:9876;192.168.100.239:9876"); /** * 设置Consumer第一次启动是从队列头部开始消费还是队列尾部开始消费<br> * 如果非第一次启动,那么按照上次消费的位置继续消费 */ consumer.setConsumeFromWhere(ConsumeFromWhere.CONSUME_FROM_FIRST_OFFSET); consumer.subscribe("TopicOrderTest", "*"); /** * 实现了MessageListenerOrderly表示一个队列只会被一个线程取到 *,第二个线程无法访问这个队列 */ consumer.registerMessageListener(new MessageListenerOrderly() { AtomicLong consumeTimes = new AtomicLong(0); public ConsumeOrderlyStatus consumeMessage(List<MessageExt> msgs, ConsumeOrderlyContext context) { // 设置自动提交 context.setAutoCommit(true); for (MessageExt msg : msgs) { System.out.println(msg + ",内容:" + new String(msg.getBody())); } try { TimeUnit.SECONDS.sleep(5L); } catch (InterruptedException e) { e.printStackTrace(); } ; return ConsumeOrderlyStatus.SUCCESS; } }); consumer.start(); System.out.println("Consumer2 Started."); } }
先启动Consumer1和Consumer2,然后启动Producer,Producer会发送15条消息
Consumer1消费情况如图,都按照顺序执行了
Consumer2消费情况如图,都按照顺序执行了
二、事务消费
这里说的主要是分布式事物。下面的例子的数据库分别安装在不同的节点上。
事物消费需要先说说什么是事务。比如说:我们跨行转账,从工商银行转到建设银行,也就是我从工商银行扣除1000元之后,我的建设银行也必须加1000元。这样才能保证数据的一致性。假如工商银行转1000元之后,建设银行的服务器突然宕机,那么我扣除了1000,但是并没有在建设银行给我加1000,就出现了数据的不一致。因此加1000和减1000才行,减1000和减1000必须一起成功,一起失败。
再比如,我们进行网购的时候,我们下单之后,订单提交成功,仓库商品的数量必须减一。但是订单可能是一个数据库,仓库数量可能又是在另个数据库里面。有可能订单提交成功之后,仓库数量服务器突然宕机。这样也出现了数据不一致的问题。
使用消息队列来解决分布式事物:
现在我们去外面饭店吃饭,很多时候都不会直接给了钱之后直接在付款的窗口递饭菜,而是付款之后他会给你一张小票,你拿着这个小票去出饭的窗口取饭。这里和我们的系统类似,提高了吞吐量。即使你到第二个窗口,师傅告诉你已经没饭了,你可以拿着这个凭证去退款,即使中途由于出了意外你无法到达窗口进行取饭,但是只要凭证还在,可以将钱退给你。这样就保证了数据的一致性。
如何保证凭证(消息)有2种方法:
1、在工商银行扣款的时候,余额表扣除1000,同时记录日志,而且这2个表是在同一个数据库实例中,可以使用本地事物解决。然后我们通知建设银行需要加1000给该用户,建设银行收到之后给我返回已经加了1000给用户的确认信息之后,我再标记日志表里面的日志为已经完成。
2、通过消息中间件
原文地址:http://www.jianshu.com/p/453c6e7ff81c
RocketMQ第一阶段发送Prepared消息时,会拿到消息的地址,第二阶段执行本地事物,第三阶段通过第一阶段拿到的地址去访问消息,并修改消息的状态。
细心的你可能又发现问题了,如果确认消息发送失败了怎么办?RocketMQ会定期扫描消息集群中的事物消息,如果发现了Prepared消息,它会向消息发送端(生产者)确认,Bob的钱到底是减了还是没减呢?如果减了是回滚还是继续发送确认消息呢?RocketMQ会根据发送端设置的策略来决定是回滚还是继续发送确认消息。这样就保证了消息发送与本地事务同时成功或同时失败。
例子:
Consumer.java
package transaction; import java.util.List; import com.alibaba.rocketmq.client.consumer.DefaultMQPushConsumer; import com.alibaba.rocketmq.client.consumer.listener.ConsumeConcurrentlyContext; import com.alibaba.rocketmq.client.consumer.listener.ConsumeConcurrentlyStatus; import com.alibaba.rocketmq.client.consumer.listener.MessageListenerConcurrently; import com.alibaba.rocketmq.client.exception.MQClientException; import com.alibaba.rocketmq.common.consumer.ConsumeFromWhere; import com.alibaba.rocketmq.common.message.MessageExt; /** * Consumer,订阅消息 */ public class Consumer { public static void main(String[] args) throws InterruptedException, MQClientException { DefaultMQPushConsumer consumer = new DefaultMQPushConsumer("transaction_Consumer"); consumer.setNamesrvAddr("192.168.100.145:9876;192.168.100.146:9876;192.168.100.149:9876;192.168.100.239:9876"); consumer.setConsumeMessageBatchMaxSize(10); /** * 设置Consumer第一次启动是从队列头部开始消费还是队列尾部开始消费<br> * 如果非第一次启动,那么按照上次消费的位置继续消费 */ consumer.setConsumeFromWhere(ConsumeFromWhere.CONSUME_FROM_FIRST_OFFSET); consumer.subscribe("TopicTransactionTest", "*"); consumer.registerMessageListener(new MessageListenerConcurrently() { public ConsumeConcurrentlyStatus consumeMessage(List<MessageExt> msgs, ConsumeConcurrentlyContext context) { try { for (MessageExt msg : msgs) { System.out.println(msg + ",内容:" + new String(msg.getBody())); } } catch (Exception e) { e.printStackTrace(); return ConsumeConcurrentlyStatus.RECONSUME_LATER;// 重试 } return ConsumeConcurrentlyStatus.CONSUME_SUCCESS;// 成功 } }); consumer.start(); System.out.println("transaction_Consumer Started."); } }
Producer.java
package transaction; import com.alibaba.rocketmq.client.exception.MQClientException; import com.alibaba.rocketmq.client.producer.SendResult; import com.alibaba.rocketmq.client.producer.TransactionCheckListener; import com.alibaba.rocketmq.client.producer.TransactionMQProducer; import com.alibaba.rocketmq.common.message.Message; /** * 发送事务消息例子 * */ public class Producer { public static void main(String[] args) throws MQClientException, InterruptedException { TransactionCheckListener transactionCheckListener = new TransactionCheckListenerImpl(); TransactionMQProducer producer = new TransactionMQProducer("transaction_Producer"); producer.setNamesrvAddr("192.168.100.145:9876;192.168.100.146:9876;192.168.100.149:9876;192.168.100.239:9876"); // 事务回查最小并发数 producer.setCheckThreadPoolMinSize(2); // 事务回查最大并发数 producer.setCheckThreadPoolMaxSize(2); // 队列数 producer.setCheckRequestHoldMax(2000); producer.setTransactionCheckListener(transactionCheckListener); producer.start(); // String[] tags = new String[] { "TagA", "TagB", "TagC", "TagD", "TagE" // }; TransactionExecuterImpl tranExecuter = new TransactionExecuterImpl(); for (int i = 1; i <= 2; i++) { try { Message msg = new Message("TopicTransactionTest", "transaction" + i, "KEY" + i, ("Hello RocketMQ " + i).getBytes()); SendResult sendResult = producer.sendMessageInTransaction(msg, tranExecuter, null); System.out.println(sendResult); Thread.sleep(10); } catch (MQClientException e) { e.printStackTrace(); } } for (int i = 0; i < 100000; i++) { Thread.sleep(1000); } producer.shutdown(); } }
TransactionExecuterImpl .java --执行本地事务
package transaction; import com.alibaba.rocketmq.client.producer.LocalTransactionExecuter; import com.alibaba.rocketmq.client.producer.LocalTransactionState; import com.alibaba.rocketmq.common.message.Message; /** * 执行本地事务 */ public class TransactionExecuterImpl implements LocalTransactionExecuter { // private AtomicInteger transactionIndex = new AtomicInteger(1); public LocalTransactionState executeLocalTransactionBranch(final Message msg, final Object arg) { System.out.println("执行本地事务msg = " + new String(msg.getBody())); System.out.println("执行本地事务arg = " + arg); String tags = msg.getTags(); if (tags.equals("transaction2")) { System.out.println("======我的操作============,失败了 -进行ROLLBACK"); return LocalTransactionState.ROLLBACK_MESSAGE; } return LocalTransactionState.COMMIT_MESSAGE; // return LocalTransactionState.UNKNOW; } }
TransactionCheckListenerImpl--未决事务,服务器回查客户端(目前已经被阉割啦)
package transaction; import com.alibaba.rocketmq.client.producer.LocalTransactionState; import com.alibaba.rocketmq.client.producer.TransactionCheckListener; import com.alibaba.rocketmq.common.message.MessageExt; /** * 未决事务,服务器回查客户端 */ public class TransactionCheckListenerImpl implements TransactionCheckListener { // private AtomicInteger transactionIndex = new AtomicInteger(0); //在这里,我们可以根据由MQ回传的key去数据库查询,这条数据到底是成功了还是失败了。 public LocalTransactionState checkLocalTransactionState(MessageExt msg) { System.out.println("未决事务,服务器回查客户端msg =" + new String(msg.getBody().toString())); // return LocalTransactionState.ROLLBACK_MESSAGE; return LocalTransactionState.COMMIT_MESSAGE; // return LocalTransactionState.UNKNOW; } }
producer端:发送数据到MQ,并且处理本地事物。这里模拟了一个成功一个失败。Consumer只会接收到本地事物成功的数据。第二个数据失败了,不会被消费。
Consumer只会接收到一个,第二个数据不会被接收到