读写分离注解

　　以前写过读写分离，今天完善成文档。

 一：概述

1.结构文档

　　

2.思路

　　组装好各个数据源，然后通过注解选择使用读或者写的数据源，将其使用AbstractRoutingDataSource中的方法determineCurrentLookuoKey进行选择datasource的key。

　　然后，通过key，就找到了要使用的数据源。

　　在数据源的这里，最后配置上连接池。

3.说明

　　本配置直接使用一个公共的连接池配置，如果需要，自己进行配置

二：程序说明

1.连接池配置

package com.jun.webpro.common.config.dataSource.properties;

import lombok.Data;
import org.springframework.boot.context.properties.ConfigurationProperties;
import org.springframework.stereotype.Component;

import java.util.List;
import java.util.Map;

/**
 * 数据库连接池配置
 */
@ConfigurationProperties(prefix = "spring.druid")
@Component
@Data
public class DruidProperties {
    /**
     * 初始化大小
     */
    private int initialSize;
    /**
     * 最小
     */
    private int minIdle;
    /**
     * 最大
     */
    private int maxActive;
    /**
     * 获取连接等待超时的时间
     */
    private int maxWait;
    /**
     * 间隔多久才进行一次检测，检测需要关闭的空闲连接，单位是毫秒
     */
    private int timeBetweenEvictionRunsMillis;
    /**
     * 一个连接在池中最小生存的时间，单位是毫秒
     */
    private int minEvictableIdleTimeMillis;

    private String validationQuery;

    private boolean testWhileIdle;

    private boolean testOnBorrow;

    private boolean testOnReturn;
    /**
     * 打开PSCache
     */
    private boolean poolPreparedStatements;
    /**
     * 并且指定每个连接上PSCache的大小
     */
    private int maxPoolPreparedStatementPerConnectionSize;
    /**
     * 监控统计拦截的filters，去掉后监控界面sql无法统计，'wall'用于防火墙
     */
    private String filters;
    /**
     * 物理连接初始化的时候执行的sql
     */
    private List<String> connectionInitSqls;
    /**
     * connectProperties属性来打开mergeSql功能；慢SQL记录
     */
    private Map<String, String> connectionProperties;
}

　　

2.主从库的配置项接口

　　因为通过这个接口，进行下面的注入

package com.jun.webpro.common.config.dataSource.properties;


public interface DataSourceProperties {

    String getDriverClassName();

    String getUrl();

    String getUsername();

    String getPassword();
}

　

3.从库配置

package com.jun.webpro.common.config.dataSource.properties;

import lombok.Data;
import lombok.Getter;
import org.springframework.boot.context.properties.ConfigurationProperties;
import org.springframework.stereotype.Component;

@Component
@Data
@Getter
@ConfigurationProperties("mysql.datasource.slave")
public class SlaveDataSourceProperties implements DataSourceProperties{
    /**
     * url
     */
    private String url;

    /**
     * driverClassName
     */
    private String driverClassName;

    /**
     * username
     */
    private String username;

    /**
     * password
     */
    private String password;
}

　　

4.主库配置

package com.jun.webpro.common.config.dataSource.properties;

import lombok.Data;
import org.springframework.boot.context.properties.ConfigurationProperties;
import org.springframework.stereotype.Component;

@Component
@Data
@ConfigurationProperties("mysql.datasource.master")
public class MasterDataSourceProperties implements DataSourceProperties{
    /**
     * url
     */
    private String url;

    /**
     * driverClassName
     */
    private String driverClassName;

    /**
     * username
     */
    private String username;

    /**
     * password
     */
    private String password;
}

　　

5.数据池配置项

package com.jun.webpro.common.config.dataSource.config;

import com.alibaba.druid.pool.DruidDataSource;
import com.jun.webpro.common.config.dataSource.properties.DataSourceProperties;
import com.jun.webpro.common.config.dataSource.properties.DruidProperties;
import org.springframework.context.annotation.Configuration;

import javax.annotation.Resource;
import java.sql.SQLException;
import java.util.Properties;

@Configuration
public class DataSourceConfig {
    // 导入数据库连接池配置
    @Resource
    private DruidProperties druidProperties;

    /**
     * @param dataSourceProperties 其他地方传参
     */
    DruidDataSource initDruidDataSource(DataSourceProperties dataSourceProperties) throws SQLException {
        try (DruidDataSource dataSource = new DruidDataSource()) {
            dataSource.setInitialSize(druidProperties.getInitialSize());
            dataSource.setMinIdle(druidProperties.getMinIdle());
            dataSource.setMaxActive(druidProperties.getMaxActive());
            dataSource.setMaxWait(druidProperties.getMaxWait());
            dataSource.setTimeBetweenEvictionRunsMillis(druidProperties.getTimeBetweenEvictionRunsMillis());
            dataSource.setMinEvictableIdleTimeMillis(druidProperties.getMinEvictableIdleTimeMillis());
            dataSource.setValidationQuery(druidProperties.getValidationQuery());
            dataSource.setTestWhileIdle(druidProperties.isTestWhileIdle());
            dataSource.setTestOnBorrow(druidProperties.isTestOnBorrow());
            dataSource.setTestOnReturn(druidProperties.isTestOnReturn());
            dataSource.setPoolPreparedStatements(druidProperties.isPoolPreparedStatements());
            dataSource.setMaxPoolPreparedStatementPerConnectionSize(druidProperties.getMaxPoolPreparedStatementPerConnectionSize());
            dataSource.setConnectionInitSqls(druidProperties.getConnectionInitSqls());
            dataSource.setFilters(druidProperties.getFilters());
            Properties properties = new Properties();
            for (String key : druidProperties.getConnectionProperties().keySet()) {
                properties.setProperty(key, druidProperties.getConnectionProperties().get(key));
            }
            dataSource.setConnectProperties(properties);
            dataSource.setDriverClassName(dataSourceProperties.getDriverClassName());
            dataSource.setUrl(dataSourceProperties.getUrl());
            dataSource.setUsername(dataSourceProperties.getUsername());
            dataSource.setPassword(dataSourceProperties.getPassword());
            return dataSource;
        }
    }
}

　　

6.主数据库配置项

package com.jun.webpro.common.config.dataSource.config;

import com.jun.webpro.common.config.dataSource.properties.MasterDataSourceProperties;
import lombok.extern.slf4j.Slf4j;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;
import org.springframework.context.annotation.Primary;
import javax.annotation.Resource;
import javax.sql.DataSource;
import java.sql.SQLException;


@Slf4j
@Configuration
public class MasterDataSourceConfig extends DataSourceConfig{
    @Resource
    private MasterDataSourceProperties dataSourceProperties;

    /**
     * 数据源
     */
    @Bean(name = "masterDataSource")
    @Primary
    public DataSource masterDataSource() throws SQLException {
        return initDruidDataSource(dataSourceProperties);
    }

}

　　

7.从数据库配置

package com.jun.webpro.common.config.dataSource.config;

import com.jun.webpro.common.config.dataSource.properties.MasterDataSourceProperties;
import com.jun.webpro.common.config.dataSource.properties.SlaveDataSourceProperties;
import lombok.extern.slf4j.Slf4j;
import org.apache.ibatis.session.SqlSessionFactory;
import org.mybatis.spring.SqlSessionFactoryBean;
import org.mybatis.spring.annotation.MapperScan;
import org.springframework.beans.factory.annotation.Qualifier;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;
import org.springframework.context.annotation.Primary;
import org.springframework.core.io.support.PathMatchingResourcePatternResolver;
import org.springframework.jdbc.datasource.DataSourceTransactionManager;

import javax.annotation.Resource;
import javax.sql.DataSource;
import java.sql.SQLException;


@Slf4j
@Configuration
public class SlaveDataSourceConfig extends DataSourceConfig{
    @Resource
    private SlaveDataSourceProperties dataSourceProperties;

    /**
     * 数据源
     */
    @Bean(name = "slaveDataSource")
    public DataSource masterDataSource() throws SQLException {
        return initDruidDataSource(dataSourceProperties);
    }
}

　　

8.mybatis配置项

　　这里用于mysql事务

package com.jun.webpro.common.config.dataSource.config;

import org.springframework.boot.context.properties.ConfigurationProperties;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;

@Configuration
public class MybatisConfig {
    @Bean("mybatis-config")
    @ConfigurationProperties(prefix = "mybatis.configuration")
    public org.apache.ibatis.session.Configuration globalConfiguration() {
        return new org.apache.ibatis.session.Configuration();
    }
}

　　

9.将数据源的key设置进ThreadLocal中

　　这里主要是注解中，将值添加进来，然后后面使用。

package com.jun.webpro.common.config.dataSource.route;

import lombok.extern.slf4j.Slf4j;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

/**
 * Description 这里切换读/写模式
 * 原理是利用ThreadLocal保存当前线程是否处于读模式（通过开始READ_ONLY注解在开始操作前设置模式为读模式，
 * 操作结束后清除该数据，避免内存泄漏，同时也为了后续在该线程进行写操作时任然为读模式
 */
@Slf4j
public class DbContextHolder {

    public static final String MASTER = "masterDataSource";
    public static final String SLAVE = "slaveDataSource";

    private static ThreadLocal<String> contextHolder= new ThreadLocal<>();

    public static void setDbType(String dbType) {
        if (dbType == null) {
            log.error("dbType为空");
            throw new NullPointerException();
        }
        log.info("设置dbType为：{}",dbType);
        contextHolder.set(dbType);
    }

    public static String getDbType() {
        return contextHolder.get();
    }

    public static void clearDbType() {
        contextHolder.remove();
    }
}

　　

10.去ThreadLocal中获取荣国注解加入的key

　　这个key可以决定走哪个库。

package com.jun.webpro.common.config.dataSource.route;

import com.jun.webpro.common.config.dataSource.route.DbContextHolder;
import com.jun.webpro.common.units.NumberUtils;
import lombok.extern.slf4j.Slf4j;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.beans.factory.annotation.Value;
import org.springframework.jdbc.datasource.lookup.AbstractRoutingDataSource;

import java.util.Objects;

/**
 * Description
 *
 */
@Slf4j
public class RoutingDataSource extends AbstractRoutingDataSource {

    @Value("${mysql.datasource.num}")
    private int num;

    @Override
    protected Object determineCurrentLookupKey() {
        String typeKey = DbContextHolder.getDbType();
        if (Objects.equals(DbContextHolder.MASTER, typeKey)) {
            log.info("使用了写库");
            return typeKey;
        }else {
            int sum = NumberUtils.getRandom(1, num);
            log.info("使用了读库{}", sum);
            return DbContextHolder.SLAVE + sum;
        }
    }
}

　　

11.读写配置，主要点是重写routingDataSource

package com.jun.webpro.common.config.dataSource.route;

import com.jun.webpro.common.config.dataSource.config.DataSourceConfig;
import com.jun.webpro.common.config.dataSource.route.DbContextHolder;
import com.jun.webpro.common.config.dataSource.route.RoutingDataSource;
import org.apache.ibatis.session.SqlSessionFactory;
import org.mybatis.spring.SqlSessionFactoryBean;
import org.mybatis.spring.annotation.MapperScan;
import org.springframework.beans.factory.annotation.Qualifier;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;
import org.springframework.core.io.support.PathMatchingResourcePatternResolver;
import org.springframework.jdbc.datasource.DataSourceTransactionManager;
import org.springframework.jdbc.datasource.lookup.AbstractRoutingDataSource;
import javax.annotation.Resource;
import javax.sql.DataSource;
import java.util.HashMap;
import java.util.Map;

/**
 * Description
 *
 */
@Configuration
@MapperScan(basePackages = "com.jun.webpro.common.domain.mapper", sqlSessionFactoryRef = "sqlSessionFactory")
public class WriteOrReadDatabaseConfig extends DataSourceConfig {
    @Resource
    private DataSource masterDataSource;

    @Resource
    private DataSource slaveDataSource;

    /**
     * 事务
     */
    @Bean
    public DataSourceTransactionManager masterTransactionManager() {
        return new DataSourceTransactionManager(routingDataSource());
    }

    @Bean
    public SqlSessionFactory sqlSessionFactory(@Qualifier("mybatis-config") org.apache.ibatis.session.Configuration configuration) throws Exception {
        final SqlSessionFactoryBean sessionFactory = new SqlSessionFactoryBean();
        sessionFactory.setDataSource(routingDataSource());
        sessionFactory.setConfiguration(configuration);
        sessionFactory.setMapperLocations(new PathMatchingResourcePatternResolver().getResources("classpath:mapper/*.xml"));
        return sessionFactory.getObject();
    }

    /**
     * 设置数据源路由，通过该类中的determineCurrentLookupKey决定使用哪个数据源
     * 工厂模式
     */
    @Bean
    public AbstractRoutingDataSource routingDataSource() {
        RoutingDataSource proxy = new RoutingDataSource();
        Map<Object, Object> targetDataSources = new HashMap<>(2);
        targetDataSources.put(DbContextHolder.MASTER, masterDataSource);
        targetDataSources.put(DbContextHolder.SLAVE+"1", slaveDataSource);
        proxy.setDefaultTargetDataSource(slaveDataSource);
        proxy.setTargetDataSources(targetDataSources);
        return proxy;
    }
}

　　

12.注解

package com.jun.webpro.common.aspect;

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

/**
 * Description 通过该接口注释的service使用读模式，其他使用写模式
 *
 * 接口注释只是一种办法，如果项目已经有代码了，通过注释可以不修改任何业务代码加持读写分离
 * 也可以通过切面根据方法开头来设置读写模式，例如getXXX()使用读模式，其他使用写模式
 *
 */
@Target({ElementType.METHOD,ElementType.TYPE})
@Retention(RetentionPolicy.RUNTIME)
public @interface ReadOnly {
}

　　

 13，注解实现

　　将要使用的库写进去

package com.jun.webpro.common.aspect.impl;

import com.jun.webpro.common.aspect.ReadOnly;
import com.jun.webpro.common.config.dataSource.route.DbContextHolder;
import org.aspectj.lang.ProceedingJoinPoint;
import org.aspectj.lang.annotation.Around;
import org.aspectj.lang.annotation.Aspect;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.core.Ordered;
import org.springframework.stereotype.Component;

/**
 * Description
 *
 */
@Aspect
@Component
public class ReadOnlyInterceptor implements Ordered {
    private static final Logger log= LoggerFactory.getLogger(ReadOnlyInterceptor.class);

    @Around("@annotation(readOnly)")
    public Object setRead(ProceedingJoinPoint joinPoint, ReadOnly readOnly) throws Throwable{
        try{
            // 通过注解，将值注册进去
            DbContextHolder.setDbType(DbContextHolder.SLAVE);
            return joinPoint.proceed();
        }finally {
            DbContextHolder.clearDbType();
            log.info("清除threadLocal");
        }
    }

    @Override
    public int getOrder() {
        return 0;
    }
}

　　

三：配置项

1.

mysql:
  datasource:
    #读库数目
    num: 1
    master:
      url: jdbc:mysql://47.103.25.1:3306/center?useUnicode=true&characterEncoding=utf-8&allowMultiQueries=true&useSSL=false&serverTimezone=UTC
      username: root
      password: 123456
      driver-class-name: com.mysql.cj.jdbc.Driver
    slave:
      url: jdbc:mysql://47.103.25.1:3306/center?useUnicode=true&characterEncoding=utf-8&allowMultiQueries=true&useSSL=false&serverTimezone=UTC
      username: root
      password: 123456
      driver-class-name: com.mysql.cj.jdbc.Driver

spring:
  # Redis配置, 使用了连接池
  redis:
    database: 0
    host: 106.14.25.1
    port: 6379
    password:
    jedis:
      pool:
        max-active: 20
        max-wait: -1
        max-idle: 20
        min-idle: 10
    timeout: 1000
  #druid数据库连接池配置
  druid:
    initialSize: 5
    minIdle: 5
    maxActive: 8
    maxWait: 60000
    timeBetweenEvictionRunsMillis: 60000
    minEvictableIdleTimeMillis: 300000
    validationQuery: SELECT 1
    testWhileIdle: true
    testOnBorrow: false
    testOnReturn: false
    poolPreparedStatements: true
    maxPoolPreparedStatementPerConnectionSize: 20
    connectionInitSqls: set names utf8mb4
    filters: stat
    connectionProperties:
      druid:
        stat:
          mergeSql: true
          slowSqlMillis: 1000

　　

 四：关于事务

1.DataSourceTranceManager

　　AbstractRoutingDataSource 只支持单库事务，也就是说切换数据源要在开启事务之前执行。

　　spring DataSourceTransactionManager进行事务管理，开启事务，会将数据源缓存到DataSourceTransactionObject对象中进行后续的commit rollback等事务操作。

　　将事务管理在数据持久 (Dao层) 开启，切换数据源的操作放在业务层进行操作，就可在事务开启之前顺利进行数据源切换，不会再出现切换失败了。

　　

五：缺点

1.不能动态的增加数据源

2.主从延迟问题

　　如果往主库增加数据，马上到从库读取，一般没有问题

　　但是数据量大的时候，因为有延迟，可能去查询的时候，查询不到