


How SpringBoot integrates RocketMQ transactions, broadcasts and sequential messages
Environment: springboot2.3.9RELEASE RocketMQ4.8.0
Dependency
<dependency> <groupid>org.springframework.boot</groupid> <artifactid>spring-boot-starter-web</artifactid> </dependency> <dependency> <groupid>org.apache.rocketmq</groupid> <artifactid>rocketmq-spring-boot-starter</artifactid> <version>2.2.0</version> </dependency>
Configuration file
server: port: 8080 --- rocketmq: nameServer: localhost:9876 producer: group: demo-mq
Normal message
Send
@Resource private RocketMQTemplate rocketMQTemplate ; public void send(String message) { rocketMQTemplate.convertAndSend("test-topic:tag2", MessageBuilder.withPayload(message).build()); }
Accept
@RocketMQMessageListener(topic = "test-topic", consumerGroup = "consumer01-group", selectorExpression = "tag1 || tag2") @Component public class ConsumerListener implements RocketMQListener<string> { @Override public void onMessage(String message) { System.out.println("接收到消息:" + message) ; } }</string>
Sequential messages
Send
@Resource private RocketMQTemplate rocketMQTemplate ; public void sendOrder(String topic, String message, String tags, int id) { rocketMQTemplate.asyncSendOrderly(topic + ":" + tags, MessageBuilder.withPayload(message).build(), "order-" + id, new SendCallback() { @Override public void onSuccess(SendResult sendResult) { System.err.println("msg-id: " + sendResult.getMsgId() + ": " + message +"\tqueueId: " + sendResult.getMessageQueue().getQueueId()) ; } @Override public void onException(Throwable e) { e.printStackTrace() ; } }); }
Here are messages sent to different queues based on hashkey
@RocketMQMessageListener(topic = "order-topic", consumerGroup = "consumer02-group", selectorExpression = "tag3 || tag4", consumeMode = ConsumeMode.ORDERLY) @Component public class ConsumerOrderListener implements RocketMQListener<string> { @Override public void onMessage(String message) { System.out.println(Thread.currentThread().getName() + " 接收到Order消息:" + message) ; } }</string>
consumeMode = ConsumeMode.ORDERLY, Indicates that the message mode is sequential mode, one queue and one thread.
Result
When consumeMode = ConsumeMode.CONCURRENTLY, the execution result is as follows:
Cluster/Broadcast Message mode
Sender
@Resource private RocketMQTemplate rocketMQTemplate ; public void send(String topic, String message, String tags) { rocketMQTemplate.send(topic + ":" + tags, MessageBuilder.withPayload(message).build()) ; }
Cluster message mode
Consumer
@RocketMQMessageListener(topic = "broad-topic", consumerGroup = "consumer03-group", selectorExpression = "tag6 || tag7", messageModel = MessageModel.CLUSTERING) @Component public class ConsumerBroadListener implements RocketMQListener<string> { @Override public void onMessage(String message) { System.out.println("ConsumerBroadListener1接收到消息:" + message) ; } }</string>
messageModel = MessageModel.CLUSTERING
Test
Start two services with ports 8080 and 8081
8080 service
8081 service
In the cluster message mode, each service receives a part of the message separately to achieve load balancing
Broadcast message mode
Consumer end
@RocketMQMessageListener(topic = "broad-topic", consumerGroup = "consumer03-group", selectorExpression = "tag6 || tag7", messageModel = MessageModel.BROADCASTING) @Component public class ConsumerBroadListener implements RocketMQListener<string> { @Override public void onMessage(String message) { System.out.println("ConsumerBroadListener1接收到消息:" + message) ; } }</string>
messageModel = MessageModel.BROADCASTING
Test
Start two services with ports 8080 and 8081
8080 service
8081 service
In cluster message mode, each service receives the same message.
Transaction message
The 3 statuses of RocketMQ transactions
TransactionStatus.CommitTransaction: Submit transaction message, consumers can consume this message
TransactionStatus.RollbackTransaction: Rolling back the transaction means that the message will be deleted and not allowed to be consumed.
TransactionStatus.Unknown: Intermediate status, which represents the need to check the message queue to determine the status.
RocketMQ's implementation of transaction messages is mainly divided into two stages: normal transaction sending and submission, and transaction information compensation process. The overall process is:
Normal transaction sending and submission stage
1. The producer sends a half message to MQServer (a half message refers to a message that the consumer cannot consume temporarily)
2. The server responds to the message writing result and the half message is sent successfully
3. Start executing the local transaction
4. Perform the Commit or Rollback operation according to the execution status of the local transaction
Compensation process of transaction information
1. If MQServer does not receive it for a long time The execution status of the local transaction will initiate a confirmation review operation request to the producer
2. After the producer receives the confirmation review request, it checks the execution status of the local transaction
3. According to After checking the results, execute the Commit or Rollback operation.
The compensation phase is mainly used to solve the problem of timeout or failure when the producer sends the Commit or Rollback operation.
Sender
@Resource private RocketMQTemplate rocketMQTemplate ; public void sendTx(String topic, Long id, String tags) { rocketMQTemplate.sendMessageInTransaction(topic + ":" + tags, MessageBuilder.withPayload( new Users(id, UUID.randomUUID().toString().replaceAll("-", ""))). setHeader("BID", UUID.randomUUID().toString().replaceAll("-", "")).build(), UUID.randomUUID().toString().replaceAll("-", "")) ; }
The listener corresponding to the producer
@RocketMQTransactionListener public class ProducerTxListener implements RocketMQLocalTransactionListener { @Resource private BusinessService bs ; @Override public RocketMQLocalTransactionState executeLocalTransaction(Message msg, Object arg) { // 这里执行本地的事务操作,比如保存数据。 try { // 创建一个日志记录表,将这唯一的ID存入数据库中,在下面的check方法中可以根据这个id查询是否有数据 String id = (String) msg.getHeaders().get("BID") ; Users users = new JsonMapper().readValue((byte[])msg.getPayload(), Users.class) ; System.out.println("消息内容:" + users + "\t参与数据:" + arg + "\t本次事务的唯一编号:" + id) ; bs.save(users, new UsersLog(users.getId(), id)) ; } catch (Exception e) { e.printStackTrace() ; return RocketMQLocalTransactionState.ROLLBACK ; } return RocketMQLocalTransactionState.COMMIT ; } @Override public RocketMQLocalTransactionState checkLocalTransaction(Message msg) { // 这里检查本地事务是否执行成功 String id = (String) msg.getHeaders().get("BID") ; System.out.println("执行查询ID为:" + id + " 的数据是否存在") ; UsersLog usersLog = bs.queryUsersLog(id) ; if (usersLog == null) { return RocketMQLocalTransactionState.ROLLBACK ; } return RocketMQLocalTransactionState.COMMIT ; } }
Consumer
@RocketMQMessageListener(topic = "tx-topic", consumerGroup = "consumer05-group", selectorExpression = "tag10") @Component public class ConsumerTxListener implements RocketMQListener<users> { @Override public void onMessage(Users users) { System.out.println("TX接收到消息:" + users) ; } }</users>
Service
@Transactional public boolean save(Users users, UsersLog usersLog) { usersRepository.save(users) ; usersLogRepository.save(usersLog) ; if (users.getId() == 1) { throw new RuntimeException("数据错误") ; } return true ; } public UsersLog queryUsersLog(String bid) { return usersLogRepository.findByBid(bid) ; }
Controller
@GetMapping("/tx/{id}") public Object sendTx(@PathVariable("id")Long id) { ps.sendTx("tx-topic", id, "tag10") ; return "send transaction success" ; }
Test
After calling the interface, the console output:
It can be seen from the print log that the consumer receives the message only after all the messages have been saved.
Deleting the data and then testing the ID as 1 will result in an error.
There is no data in the database. . .
Isn’t it very complicated? It can be handled in 2 stages.
The above is the detailed content of How SpringBoot integrates RocketMQ transactions, broadcasts and sequential messages. For more information, please follow other related articles on the PHP Chinese website!

Hot AI Tools

Undresser.AI Undress
AI-powered app for creating realistic nude photos

AI Clothes Remover
Online AI tool for removing clothes from photos.

Undress AI Tool
Undress images for free

Clothoff.io
AI clothes remover

Video Face Swap
Swap faces in any video effortlessly with our completely free AI face swap tool!

Hot Article

Hot Tools

Notepad++7.3.1
Easy-to-use and free code editor

SublimeText3 Chinese version
Chinese version, very easy to use

Zend Studio 13.0.1
Powerful PHP integrated development environment

Dreamweaver CS6
Visual web development tools

SublimeText3 Mac version
God-level code editing software (SublimeText3)

Hot Topics



Introduction to Jasypt Jasypt is a java library that allows a developer to add basic encryption functionality to his/her project with minimal effort and does not require a deep understanding of how encryption works. High security for one-way and two-way encryption. , standards-based encryption technology. Encrypt passwords, text, numbers, binaries... Suitable for integration into Spring-based applications, open API, for use with any JCE provider... Add the following dependency: com.github.ulisesbocchiojasypt-spring-boot-starter2. 1.1Jasypt benefits protect our system security. Even if the code is leaked, the data source can be guaranteed.

Usage scenario 1. The order was placed successfully but the payment was not made within 30 minutes. The payment timed out and the order was automatically canceled. 2. The order was signed and no evaluation was conducted for 7 days after signing. If the order times out and is not evaluated, the system defaults to a positive rating. 3. The order is placed successfully. If the merchant does not receive the order for 5 minutes, the order is cancelled. 4. The delivery times out, and push SMS reminder... For scenarios with long delays and low real-time performance, we can Use task scheduling to perform regular polling processing. For example: xxl-job Today we will pick

1. Redis implements distributed lock principle and why distributed locks are needed. Before talking about distributed locks, it is necessary to explain why distributed locks are needed. The opposite of distributed locks is stand-alone locks. When we write multi-threaded programs, we avoid data problems caused by operating a shared variable at the same time. We usually use a lock to mutually exclude the shared variables to ensure the correctness of the shared variables. Its scope of use is in the same process. If there are multiple processes that need to operate a shared resource at the same time, how can they be mutually exclusive? Today's business applications are usually microservice architecture, which also means that one application will deploy multiple processes. If multiple processes need to modify the same row of records in MySQL, in order to avoid dirty data caused by out-of-order operations, distribution needs to be introduced at this time. The style is locked. Want to achieve points

Springboot reads the file, but cannot access the latest development after packaging it into a jar package. There is a situation where springboot cannot read the file after packaging it into a jar package. The reason is that after packaging, the virtual path of the file is invalid and can only be accessed through the stream. Read. The file is under resources publicvoidtest(){Listnames=newArrayList();InputStreamReaderread=null;try{ClassPathResourceresource=newClassPathResource("name.txt");Input

SpringBoot and SpringMVC are both commonly used frameworks in Java development, but there are some obvious differences between them. This article will explore the features and uses of these two frameworks and compare their differences. First, let's learn about SpringBoot. SpringBoot was developed by the Pivotal team to simplify the creation and deployment of applications based on the Spring framework. It provides a fast, lightweight way to build stand-alone, executable

When Springboot+Mybatis-plus does not use SQL statements to perform multi-table adding operations, the problems I encountered are decomposed by simulating thinking in the test environment: Create a BrandDTO object with parameters to simulate passing parameters to the background. We all know that it is extremely difficult to perform multi-table operations in Mybatis-plus. If you do not use tools such as Mybatis-plus-join, you can only configure the corresponding Mapper.xml file and configure The smelly and long ResultMap, and then write the corresponding sql statement. Although this method seems cumbersome, it is highly flexible and allows us to

1. Customize RedisTemplate1.1, RedisAPI default serialization mechanism. The API-based Redis cache implementation uses the RedisTemplate template for data caching operations. Here, open the RedisTemplate class and view the source code information of the class. publicclassRedisTemplateextendsRedisAccessorimplementsRedisOperations, BeanClassLoaderAware{//Declare key, Various serialization methods of value, the initial value is empty @NullableprivateRedisSe

In projects, some configuration information is often needed. This information may have different configurations in the test environment and the production environment, and may need to be modified later based on actual business conditions. We cannot hard-code these configurations in the code. It is best to write them in the configuration file. For example, you can write this information in the application.yml file. So, how to get or use this address in the code? There are 2 methods. Method 1: We can get the value corresponding to the key in the configuration file (application.yml) through the ${key} annotated with @Value. This method is suitable for situations where there are relatively few microservices. Method 2: In actual projects, When business is complicated, logic
