我正在研究mysql主从复制.我正在使用spring data jpa(spring boot).

我需要的是进入主服务器的所有写操作和要在多个只读从属设备之间平均分配的只读操作.

为此,我需要：

使用特殊的JDBC驱动程序：com.mysql.jdbc.ReplicationDriver

在网址中设置复制：：

spring:
    datasource:
        driverClassName: com.mysql.jdbc.ReplicationDriver
        url: jdbc:mysql:replication://127.0.0.1:3306,127.0.0.1:3307/MyForum?user=root&password=password&autoReconnect=true
        test-on-borrow: true
        validation-query: SELECT 1
    database: MYSQL

自动提交需要关闭. (*)
连接需要设置为只读.

为了确保将JDBC连接设置为只读,我创建了一个注释和一个简单的AOP拦截器.

注解

package com.xyz.forum.replication;

import java.lang.annotation.ElementType;
import java.lang.annotation.Retention;
import java.lang.annotation.RetentionPolicy;
import java.lang.annotation.Target; 

/**
 * Created by Bhupati Patel on 02/11/15.
 */
@Retention(RetentionPolicy.RUNTIME)
@Target(ElementType.METHOD)
public @interface ReadOnlyConnection {
}

拦截器

package com.xyz.forum.replication;

import org.aspectj.lang.ProceedingJoinPoint;
import org.aspectj.lang.annotation.Around;
import org.aspectj.lang.annotation.Aspect;
import org.aspectj.lang.annotation.Pointcut;
import org.hibernate.Session;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Component;

import javax.persistence.EntityManager;

/**
 * Created by Bhupati Patel on 02/11/15.
 */

@Aspect
@Component
public class ConnectionInterceptor {

    private Logger logger;

    public ConnectionInterceptor() {
        logger = LoggerFactory.getLogger(getClass());
        logger.info("ConnectionInterceptor Started");
    }

    @Autowired
    private EntityManager entityManager;

    @Pointcut("@annotation(com.xyz.forum.replication.ReadOnlyConnection)")
    public void inReadOnlyConnection(){}


    @Around("inReadOnlyConnection()")
    public Object proceed(ProceedingJoinPoint pjp) throws Throwable {
        Session session = entityManager.unwrap(Session.class);
        ConnectionReadOnly readOnlyWork = new ConnectionReadOnly();

        try{
            session.doWork(readOnlyWork);
            return pjp.proceed();
        } finally {
            readOnlyWork.switchBack();
        }
    }

}

以下是我的春季数据存储库

package com.xyz.forum.repositories;

import com.xyz.forum.entity.Topic;
import org.springframework.data.repository.Repository;

import java.util.List;

/**
 * Created by Bhupati Patel on 16/04/15.
 */
public interface TopicRepository extends Repository<Topic,Integer>{
    Topic save(Topic topic);
    Topic findByTopicIdAndIsDeletedFalse(Integer topicId);
    List<Topic> findByIsDeletedOrderByTopicOrderAsc(Boolean isDelete);

}

以下是我的Manager(Service)类.

package com.xyz.forum.manager;

import com.xyz.forum.domain.entry.impl.TopicEntry;

import com.xyz.forum.domain.exception.impl.AuthException;

import com.xyz.forum.domain.exception.impl.NotFoundException;
import com.xyz.forum.entity.Topic;
import com.xyz.forum.replication.ReadOnlyConnection;
import com.xyz.forum.repositories.TopicRepository;
import com.xyz.forum.utils.converter.TopicConverter;
import org.apache.commons.lang3.StringUtils;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.stereotype.Repository;
import org.springframework.transaction.annotation.Transactional;

import java.util.*;

/**
 * Created by Bhupati Patel on 16/04/15.
 */
@Repository
public class TopicManager {
    @Autowired
    TopicRepository topicRepository;

    @Transactional
    public TopicEntry save(TopicEntry topicEntry) {
        Topic topic = TopicConverter.fromEntryToEntity(topicEntry);
        return TopicConverter.fromEntityToEntry(topicRepository.save(topic));
    }

    @ReadOnlyConnection
    public TopicEntry get(Integer id) {
        Topic topicFromDb = topicRepository.findByTopicIdAndIsDeletedFalse(id);
        if(topicFromDb == null) {
            throw new NotFoundException("Invalid Id", "Topic Id [" + id + "] doesn't exist ");
        }
        return TopicConverter.fromEntityToEntry(topicFromDb);
    }
}

在上面的代码中,@ ReadOnlyConnection批注在管理器或服务层中指定.上面的代码对我来说很好用.在服务层中,我仅从从数据库中读取并写入主数据库中,这是一种微不足道的情况.

话虽如此,我的实际要求是我应该能够在存储库级别本身中使用@ReadOnlyConnection,因为我有很多业务逻辑要在其他服务层类别中进行读/写操作,因此不能将@ReadOnlyConnection放进去.在服务层.

我应该可以使用这样的东西

public interface TopicRepository extends Repository<Topic,Integer>{
    Topic save(Topic topic);
    @ReadOnlyConnection
    Topic findByTopicIdAndIsDeletedFalse(Integer topicId);
    @ReadOnlyConnection
    List<Topic> findByIsDeletedOrderByTopicOrderAsc(Boolean isDelete);

}

就像spring的@Transactional或@Modifying或@Query注释一样.以下是我所指的示例.

    public interface AnswerRepository extends Repository<Answer,Integer> {
    @Transactional
    Answer save(Answer answer);

    @Transactional
    @Modifying
    @Query("update Answer ans set ans.isDeleted = 1, ans.deletedBy = :deletedBy, ans.deletedOn = :deletedOn " +
            "where ans.questionId = :questionId and ans.isDeleted = 0")
    void softDeleteBulkAnswers(@Param("deletedBy") String deletedBy, @Param("deletedOn") Date deletedOn,
                               @Param("questionId") Integer questionId);
}

我对Aspectj和Aop World还是新手,我在ConnectionInterceptor中尝试了很多切入点的正则表达式,但是它们都不起作用.我已经尝试了很久了,但是还没有运气.

如何完成所要求的任务.

解决方法:

我无法在方法级别使用自定义批注@ReadOnlyConnection(如@Transactional),但是我遇到了一点麻烦.

我在下面粘贴代码段.

@Aspect
@Component
@EnableAspectJAutoProxy
public class ConnectionInterceptor {

    private Logger logger;
    private static final String JPA_PREFIX = "findBy";
    private static final String CUSTOM_PREFIX = "read";

    public ConnectionInterceptor() {
        logger = LoggerFactory.getLogger(getClass());
        logger.info("ConnectionInterceptor Started");
    }

    @Autowired
    private EntityManager entityManager;

    @Pointcut("this(org.springframework.data.repository.Repository)")
    public void inRepositoryLayer() {}

    @Around("inRepositoryLayer()")
    public Object proceed(ProceedingJoinPoint pjp) throws Throwable {
        String methodName = pjp.getSignature().getName();
        if (StringUtils.startsWith(methodName, JPA_PREFIX) || StringUtils.startsWith(methodName, CUSTOM_PREFIX)) {
            System.out.println("I'm there!" );
            Session session = entityManager.unwrap(Session.class);
            ConnectionReadOnly readOnlyWork = new ConnectionReadOnly();

            try{
                session.doWork(readOnlyWork);
                return pjp.proceed();
            } finally {
                readOnlyWork.switchBack();
            }
        }
        return pjp.proceed();
    }
}

因此,在上面的代码中,我使用的切入点如下

@Pointcut("this(org.springframework.data.repository.Repository)")
public void inRepositoryLayer() {}

它的作用是

any join point (method execution only in Spring AOP) where the proxy implements the Repository interface

你可以看看
http://docs.spring.io/spring/docs/current/spring-framework-reference/html/aop.html

现在,我所有的存储库读取查询方法都以前缀“ findByXXX”(默认的spring-data-jpa可读方法)或“ readXXX”(带有@Query注释的自定义读取方法)开头,在我的around方法执行中与上述切入点匹配.根据我的要求,我将JDBC连接设置为true.

Session session = entityManager.unwrap(Session.class);
ConnectionReadOnly readOnlyWork = new ConnectionReadOnly();

和我的ConnectionReadOnly看起来像下面

package com.xyz.forum.replication;

import org.hibernate.jdbc.Work;

import java.sql.Connection;
import java.sql.SQLException;

/**
 * Created by Bhupati Patel on 04/11/15.
 */
public class ConnectionReadOnly implements Work {

    private Connection connection;
    private boolean autoCommit;
    private boolean readOnly;

    @Override
    public void execute(Connection connection) throws SQLException {
        this.connection = connection;
        this.autoCommit = connection.getAutoCommit();
        this.readOnly = connection.isReadOnly();
        connection.setAutoCommit(false);
        connection.setReadOnly(true);
    }

    //method to restore the connection state before intercepted
    public void switchBack() throws SQLException{
        connection.setAutoCommit(autoCommit);
        connection.setReadOnly(readOnly);
    }
}

因此,以上设置可以满足我的要求.