Spring Boot2 (IV): Spring Boot using separate read and write multiple data sources to achieve

Foreword

The actual business scenarios, can not be just a library, so there is a sub-library sub-table, there is more data sources. To achieve a separate read and write, the main library is responsible for additions and changes to delete, query in charge from the library. This article will realize how Spring Boot multi-source data, dynamic data source switching, separate read and write operations.

Code deployment

Quick new projects spring-boot project

1, add maven dependence

<dependency>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter-web</artifactId>
</dependency>

<dependency>
    <groupId>org.mybatis.spring.boot</groupId>
    <artifactId>mybatis-spring-boot-starter</artifactId>
    <version>2.0.1</version>
</dependency>

<dependency>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter-test</artifactId>
    <scope>test</scope>
</dependency>

<dependency>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter</artifactId>
</dependency>

<dependency>
    <groupId>mysql</groupId>
    <artifactId>mysql-connector-java</artifactId>
    <scope>runtime</scope>
</dependency>

<dependency>
    <groupId>org.projectlombok</groupId>
    <artifactId>lombok</artifactId>
    <optional>true</optional>
</dependency>

2, application to configure multiple data sources to read configuration

Before and tutorials, first configure application.yml

#指定配置文件为test
spring:
  profiles:
    active: test

#配置Mybatis
mybatis:
  configuration:
    # 开启驼峰命名转换，如：Table(create_time) -> Entity(createTime)。不需要我们关心怎么进行字段匹配，mybatis会自动识别`大写字母与下划线`
    map-underscore-to-camel-case: true

#打印SQL日志
logging:
  level:
    com.niaobulashi.mapper.*: DEBUG

Wherein this print log SQL, because multiple data sources, the following distinguishing different packet mapper source database xml file, indicated by *.

Arranged as application-test.yml

spring:
  datasource:
    #主库
    master:
      jdbc-url: jdbc:mysql://127.0.0.1:3306/test?serverTimezone=UTC&useUnicode=true&characterEncoding=utf-8&useSSL=true
      username: root
      password: root
      driver-class-name: com.mysql.cj.jdbc.Driver
    #从库
    slave:
      jdbc-url: jdbc:mysql://127.0.0.1:3306/test2?serverTimezone=UTC&useUnicode=true&characterEncoding=utf-8&useSSL=true
      username: root
      password: root
      driver-class-name: com.mysql.cj.jdbc.Driver

Starting spring.datasource node, distinguishing main library master, from the library slave. Database repository connected to the main test, the database connection from library test2.

Note : It should be noted that, starting from the Spring Boot2, some configuration changed when configuring multiple data sources, many online tutorials using spring.datasource.url. There will be jdbcUrl is required with driverClassName.problems.

Solution : When configuring multiple data sources, a spring.datasource.urlconfiguration tospring.datasource.jdbc-url

3, adding the main library configuration information

Based on well-known blogger: pure smile, write blog posts we have to analyze wave

First, look at the code of the main library configurations:

@Configuration
@MapperScan(basePackages = "com.niaobulashi.mapper.master", sqlSessionTemplateRef = "masterSqlSessionTemplate")
public class DataSourceMasterConfig {

    /**
     * 是application-test.yml中的spring.datasource.master配置生效
     * @return
     */
    @Bean(name = "masterDataSource")
    @ConfigurationProperties(prefix = "spring.datasource.master")
    @Primary
    public DataSource masterDataSource() {
        return DataSourceBuilder.create().build();
    }

    /**
     * 将配置信息注入到SqlSessionFactoryBean中
     * @param dataSource    数据库连接信息
     * @return
     * @throws Exception
     */
    @Bean(name = "masterSqlSessionFactory")
    @Primary
    public SqlSessionFactory masterSqlSessionFactory(@Qualifier("masterDataSource") DataSource dataSource) throws Exception {
        SqlSessionFactoryBean bean = new SqlSessionFactoryBean();
        bean.setDataSource(dataSource);
        bean.setMapperLocations(new PathMatchingResourcePatternResolver().getResources("classpath:mapper/master/*.xml"));
        return bean.getObject();
    }

    /**
     * 事务管理器，在实例化时注入主库master
     * @param dataSource
     * @return
     */
    @Bean(name = "masterTransactionManager")
    @Primary
    public DataSourceTransactionManager masterTransactionManager(@Qualifier("masterDataSource") DataSource dataSource) {
        return new DataSourceTransactionManager(dataSource);
    }

    /**
     * SqlSessionTemplate具有线程安全性
     * @param sqlSessionFactory
     * @return
     * @throws Exception
     */
    @Bean(name = "masterSqlSessionTemplate")
    @Primary
    public SqlSessionTemplate masterSqlSessionTemplate(@Qualifier("masterSqlSessionFactory") SqlSessionFactory sqlSessionFactory) throws Exception {
        return new SqlSessionTemplate(sqlSessionFactory);
    }
}

Problem : see this masterSqlSessionFactory, SqlSessionFactoryBeanjust get the spring.datasource.masterdatabase connection information, and did not obtain the configuration information database of more than mybatis.configurationcause we need to configure hump naming, configuration information and is not injected into SqlSessionFactoryBean. This results in a query that can not be resolved encountered underscore the appropriate fields user_id, dept_id, create_time

Solution : Add Configuration in the configuration

Meanwhile, the configuration information is injected into SqlSessionFactoryBean

/**
 * 将配置信息注入到SqlSessionFactoryBean中
 * @param dataSource    数据库连接信息
 * @return
 * @throws Exception
 */
@Bean(name = "slaveSqlSessionFactory")
public SqlSessionFactory slaveSqlSessionFactory(@Qualifier("slaveDataSource") DataSource dataSource) throws Exception {
    SqlSessionFactoryBean bean = new SqlSessionFactoryBean();
    // 使配置信息加载到类中，再注入到SqlSessionFactoryBean
    org.apache.ibatis.session.Configuration configuration = new org.apache.ibatis.session.Configuration();
    configuration.setMapUnderscoreToCamelCase(true);
    bean.setConfiguration(configuration);
    bean.setDataSource(dataSource);
    bean.setMapperLocations(new PathMatchingResourcePatternResolver().getResources("classpath:mapper/slave/*.xml"));
    return bean.getObject();
}

4, add configuration information from the library

和添加主库配置信息一样，只不过不同的是，不需要添加@Primary首选注解

代码如下

@Configuration
@MapperScan(basePackages = "com.niaobulashi.mapper.slave", sqlSessionTemplateRef = "slaveSqlSessionTemplate")
public class DataSourceSlaveConfig {

    /**
     * 是application-test.yml中的spring.datasource.master配置生效
     * @return
     */
    @Bean(name = "slaveDataSource")
    @ConfigurationProperties(prefix = "spring.datasource.slave")
    public DataSource slaveDataSource() {
        return DataSourceBuilder.create().build();
    }

    /**
     * 将配置信息注入到SqlSessionFactoryBean中
     * @param dataSource    数据库连接信息
     * @return
     * @throws Exception
     */
    @Bean(name = "slaveSqlSessionFactory")
    public SqlSessionFactory slaveSqlSessionFactory(@Qualifier("slaveDataSource") DataSource dataSource) throws Exception {
        SqlSessionFactoryBean bean = new SqlSessionFactoryBean();
        // 使配置信息加载到类中，再注入到SqlSessionFactoryBean
        org.apache.ibatis.session.Configuration configuration = new org.apache.ibatis.session.Configuration();
        configuration.setMapUnderscoreToCamelCase(true);
        bean.setConfiguration(configuration);
        bean.setDataSource(dataSource);
        bean.setMapperLocations(new PathMatchingResourcePatternResolver().getResources("classpath:mapper/slave/*.xml"));
        return bean.getObject();
    }

    /**
     * 事务管理器，在实例化时注入主库master
     * @param dataSource
     * @return
     */
    @Bean(name = "slaveTransactionManager")
    public DataSourceTransactionManager slaveTransactionManager(@Qualifier("slaveDataSource") DataSource dataSource) {
        return new DataSourceTransactionManager(dataSource);
    }

    /**
     * SqlSessionTemplate具有线程安全性
     * @param sqlSessionFactory
     * @return
     * @throws Exception
     */
    @Bean(name = "slaveSqlSessionTemplate")
    public SqlSessionTemplate slaveSqlSessionTemplate(@Qualifier("slaveSqlSessionFactory") SqlSessionFactory sqlSessionFactory) throws Exception {
        return new SqlSessionTemplate(sqlSessionFactory);
    }
}

5、扩展配置方法会报错

在网上还看到这样一种配置，单独通过@ConfigurationProperties注解配置Mybatis的配置信息如下

/**
 * 试application.yml中的mybatis.configuration配置生效，如果不主动配置，由于@Order配置顺序不同，讲导致配置不能及时生效
 * 使配置信息加载到类中，再注入到SqlSessionFactoryBean
 * @return
 */
@Bean
@ConfigurationProperties(prefix = "mybatis.configuration")
public org.apache.ibatis.session.Configuration configuration() {
    return new org.apache.ibatis.session.Configuration();
}

其中prefix，在主库和从库中的id是一样的，必须保持不同，否则idea就会提示报错Duplicate prefix

导致只有主库可以执行Mybatis的配置，从库无效。

@Bean(name = "masterSqlSessionFactory")
    @Primary
    public SqlSessionFactory masterSqlSessionFactory(@Qualifier("masterDataSource") DataSource dataSource, org.apache.ibatis.session.Configuration configuration) throws Exception {
        SqlSessionFactoryBean bean = new SqlSessionFactoryBean();
        // 使配置信息加载到类中，再注入到SqlSessionFactoryBean
        bean.setConfiguration(configuration);
        bean.setDataSource(dataSource);
        bean.setMapperLocations(new PathMatchingResourcePatternResolver().getResources("classpath:mapper/master/*.xml"));
        return bean.getObject();
    }

这块验证只有主库有效，从库的驼峰方法解析无效。后续再来研究下。。。

6、数据层代码

代码结构如下

其中SysUserMasterDao代码

public interface SysUserMasterDao {
    
    /**
     * 根据userId查询用户信息
     * @param userId  用户ID
     */
    List<SysUserEntity> queryUserInfo(Long userId);

    /**
     * 查询所有用户信息
     */
    List<SysUserEntity> queryUserAll();

    /**
     * 根据userId更新用户的邮箱和手机号
     * @return
     */
    int updateUserInfo(SysUserEntity user);

}

7、resource下数据执行语句

SysCodeMasterDao.xml

<mapper namespace="com.niaobulashi.mapper.master.SysUserMasterDao">

    <!--查询所有用户信息-->
    <select id="queryUserAll" resultType="com.niaobulashi.entity.SysUserEntity">
        SELECT
            ur.*
        FROM
            sys_user ur
        WHERE
            1 = 1
    </select>

    <!--根据用户userId查询用户信息-->
    <select id="queryUserInfo" resultType="com.niaobulashi.entity.SysUserEntity">
        SELECT
            ur.*
        FROM
            sys_user ur
        WHERE
            1 = 1
          AND ur.user_id = #{userId}
    </select>

    <!-- 根据UserId，更新邮箱和手机号 -->
    <update id="updateUserInfo" parameterType="com.niaobulashi.entity.SysUserEntity">
        UPDATE sys_user u
        <set>
            <if test="email != null">
                u.email = #{email},
            </if>
            <if test="mobile != null">
                u.mobile = #{mobile},
            </if>
        </set>
        WHERE
        u.user_id = #{userId}
    </update>

</mapper>

8、Controller层测试

@RestController
public class SysUserController {

    @Autowired
    private SysUserMasterDao sysUserMasterDao;

    @Autowired
    private SysUserSlaveDao sysUserSlaveDao;

    /**
     * 查询所有用户信息Master
     * @return
     */
    @RequestMapping("/getUserMasterAll")
    private List<SysUserEntity> getUserMaster() {
        System.out.println("查询主库");
        List<SysUserEntity> userList = sysUserMasterDao.queryUserAll();
        return userList;
    }

    /**
     * 查询所有用户信息Slave
     * @return
     */
    @RequestMapping("/getUserSlaveAll")
    private List<SysUserEntity> getUserSlave() {
        System.out.println("查询从库");
        List<SysUserEntity> userList = sysUserSlaveDao.queryUserAll();
        return userList;
    }

    /**
     * 根据userId查询用户信息Master
     * @return
     */
    @RequestMapping("/getUserMasterById")
    private List<SysUserEntity> getUserMasterById(@RequestParam(value = "userId", required = false) Long userId) {
        List<SysUserEntity> userList = sysUserMasterDao.queryUserInfo(userId);
        return userList;
    }

    /**
     * 根据userId查询用户信息Slave
     * @return
     */
    @RequestMapping("/getUserSlaveById")
    private List<SysUserEntity> getUserSlaveById(@RequestParam(value = "userId", required = false) Long userId) {
        List<SysUserEntity> userList = sysUserSlaveDao.queryUserInfo(userId);
        return userList;
    }

}

发送查询所有用户接口

主库：http://localhost:8080/getUserMasterAll

从库：http://localhost:8080/getUserSlaveAll

总结

1、通过多数据源方式实现数据库层面的读写分离

2、多数据源链接数据库是，使用spring.datasource.jdbc-url

3、多数据源的mybatis.configuration配置注意需要手动注入SqlSessionFactory

示例代码-github