多数据源配置，经常遇到的问题。
网上看到的都是别人总结的，最近一个项目，有需要用到2个数据源，就自己配置了下。
非常符合自己的思路：清晰、简洁、方便扩展。
1、Spring核心数据源和事务配置
spring-info-datasource.xml
<?xml version="1.0" encoding="UTF-8"?>
<beans xmlns="http://www.springframework.org/schema/beans"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:context="http://www.springframework.org/schema/context"
xmlns:jdbc="http://www.springframework.org/schema/jdbc" xmlns:jee="http://www.springframework.org/schema/jee"
xmlns:tx="http://www.springframework.org/schema/tx" xmlns:util="http://www.springframework.org/schema/util"
xmlns:task="http://www.springframework.org/schema/task" xmlns:aop="http://www.springframework.org/schema/aop"
xsi:schemaLocation="
http://www.springframework.org/schema/beans http://www.springframework.org/schema/beans/spring-beans-3.0.xsd
http://www.springframework.org/schema/context http://www.springframework.org/schema/context/spring-context-3.0.xsd
http://www.springframework.org/schema/jdbc http://www.springframework.org/schema/jdbc/spring-jdbc-3.0.xsd
http://www.springframework.org/schema/jee http://www.springframework.org/schema/jee/spring-jee-3.0.xsd
http://www.springframework.org/schema/tx http://www.springframework.org/schema/tx/spring-tx-3.0.xsd
http://www.springframework.org/schema/util http://www.springframework.org/schema/util/spring-util-3.0.xsd
http://www.springframework.org/schema/task http://www.springframework.org/schema/task/spring-task-3.0.xsd
http://www.springframework.org/schema/aop http://www.springframework.org/schema/aop/spring-aop-3.0.xsd">
<!-- 多数据源，根据key动态切换 -->
<bean id="lifecycleDataSource" class="com.alibaba.druid.pool.DruidDataSource"
destroy-method="close">
<!-- 数据库基本信息配置 -->
<property name="driverClassName" value="${lifecycleDatasourceAsset.driverClassName}" />
<property name="url" value="${lifecycleDatasourceAsset.url}" />
<property name="username" value="${lifecycleDatasourceAsset.username}" />
<property name="password" value="${lifecycleDatasourceAsset.password}" />
<!-- 最大并发连接数 -->
<property name="maxActive" value="${lifecycleDatasourceAsset.maxActive}" />
<!-- 初始化连接数量 -->
<property name="initialSize" value="${lifecycleDatasourceAsset.initialSize}" />
<!-- 配置获取连接等待超时的时间 -->
<property name="maxWait" value="${lifecycleDatasourceAsset.maxWait}" />
<!-- 最小空闲连接数 -->
<property name="minIdle" value="${lifecycleDatasourceAsset.minIdle}" />
<!-- 配置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 -->
<property name="timeBetweenEvictionRunsMillis"
value="${lifecycleDatasourceAsset.timeBetweenEvictionRunsMillis}" />
<!-- 配置一个连接在池中最小生存的时间，单位是毫秒 -->
<property name="minEvictableIdleTimeMillis"
value="${lifecycleDatasourceAsset.minEvictableIdleTimeMillis}" />
<property name="validationQuery" value="SELECT 1">
</property>
<!-- 打开removeAbandoned功能 -->
<property name="removeAbandoned" value="true" />
<!-- 1800秒，也就是30分钟 -->
<property name="removeAbandonedTimeout" value="180" />
</bean>
<bean id="jdbuyDataSource" class="com.alibaba.druid.pool.DruidDataSource"
destroy-method="close">
<!-- 数据库基本信息配置 -->
<property name="driverClassName" value="${jdbuyDatasourceAsset.driverClassName}" />
<property name="url" value="${jdbuyDatasourceAsset.url}" />
<property name="username" value="${jdbuyDatasourceAsset.username}" />
<property name="password" value="${jdbuyDatasourceAsset.password}" />
<!-- 最大并发连接数 -->
<property name="maxActive" value="${jdbuyDatasourceAsset.maxActive}" />
<!-- 初始化连接数量 -->
<property name="initialSize" value="${jdbuyDatasourceAsset.initialSize}" />
<!-- 配置获取连接等待超时的时间 -->
<property name="maxWait" value="${jdbuyDatasourceAsset.maxWait}" />
<!-- 最小空闲连接数 -->
<property name="minIdle" value="${jdbuyDatasourceAsset.minIdle}" />
<!-- 配置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 -->
<property name="timeBetweenEvictionRunsMillis"
value="${jdbuyDatasourceAsset.timeBetweenEvictionRunsMillis}" />
<!-- 配置一个连接在池中最小生存的时间，单位是毫秒 -->
<property name="minEvictableIdleTimeMillis"
value="${jdbuyDatasourceAsset.minEvictableIdleTimeMillis}" />
<property name="validationQuery" value="SELECT 1">
</property>
<!-- 打开removeAbandoned功能 -->
<property name="removeAbandoned" value="true" />
<!-- 1800秒，也就是30分钟 -->
<property name="removeAbandonedTimeout" value="180" />
</bean>
<bean id="multiDataSource" class="com.jd.zc.lifecycle.dao.datasource.MultiDataSource">
<property name="defaultTargetDataSource" ref="lifecycleDataSource"/>
<property name="targetDataSources">
<map>
<entry key="jdbuyDataSource" value-ref="jdbuyDataSource"/>
<entry key="lifecycleataSource" value-ref="lifecycleDataSource"/>
</map>
</property>
</bean>
<!-- 事务，每个数据源配置单独的事务 -->
<bean name="lifecycleTransactionManager"
class="org.springframework.jdbc.datasource.DataSourceTransactionManager">
<property name="dataSource" ref="lifecycleDataSource" />
<qualifier value="lifecycle"/>
</bean>
<tx:annotation-driven transaction-manager="lifecycleTransactionManager"
proxy-target-class="true" />
<bean name="jdbuyTransactionManager"
class="org.springframework.jdbc.datasource.DataSourceTransactionManager">
<property name="dataSource" ref="jdbuyDataSource" />
<qualifier value="jdbuy"/>
</bean>
<!-- 注解事务 -->
<tx:annotation-driven transaction-manager="jdbuyTransactionManager"
proxy-target-class="true" />
<!-- 配置mybatis -->
<bean id="sqlSessionFactory" class="org.mybatis.spring.SqlSessionFactoryBean">
<property name="dataSource" ref="multiDataSource" />
<property name="configLocation" value="classpath:mybatis/mybatis-config.xml" />
<!-- mapper扫描 -->
<property name="mapperLocations">
<value>classpath:mybatis/mapper/**/*Mapper.xml</value>
</property>
</bean>
<!--创建数据映射器，数据映射器必须为接口 -->
<bean class="org.mybatis.spring.mapper.MapperScannerConfigurer">
<property name="annotationClass" value="org.springframework.stereotype.Repository" />
<property name="basePackage" value="com.jd.zc.lifecycle.dao" />
</bean>
<bean id="sqlSessionTemplate" class="org.mybatis.spring.SqlSessionTemplate">
<constructor-arg ref="sqlSessionFactory" />
</bean>
<context:component-scan base-package="com.jd.zc.lifecycle.dao" />
</beans>数据源属性文件jdbc.properties
#lifecycle
lifecycleDatasourceAsset.driverClassName=com.mysql.jdbc.Driver
lifecycleDatasourceAsset.url=jdbc:mysql://lifecycle?characterEncoding=UTF-8&allowMultiQueries=true
lifecycleDatasourceAsset.initialSize=0
lifecycleDatasourceAsset.maxActive=20
lifecycleDatasourceAsset.minIdle=2
lifecycleDatasourceAsset.maxWait=1000
lifecycleDatasourceAsset.minEvictableIdleTimeMillis=300000
lifecycleDatasourceAsset.timeBetweenEvictionRunsMillis=120000
#jdbuy
jdbuyDatasourceAsset.driverClassName=com.mysql.jdbc.Driver
jdbuyDatasourceAsset.url=jdbc:mysql://192.168./jdbuy?characterEncoding=UTF-8&allowMultiQueries=true
jdbuyDatasourceAsset.initialSize=0
jdbuyDatasourceAsset.maxActive=20
jdbuyDatasourceAsset.minIdle=2
jdbuyDatasourceAsset.maxWait=1000
jdbuyDatasourceAsset.minEvictableIdleTimeMillis=300000
jdbuyDatasourceAsset.timeBetweenEvictionRunsMillis=1200002、Mapper映射xml文件
具体到某个Mapper文件，都是很标准正常的配置。
出于方便好维护，每个数据源的mapper文件单独放在一个目录。
3、Mapper接口类
BaseMapper：通用接口
public interface BaseMapper<ID, Entity, Bean> {
// read
public Entity get(ID id);
// write
public int add(Entity entity);
public int update(Entity entity);
public int remove(ID id);
public int removeByIdList(List<ID> idList);
}
public interface JdbuyBaseMapper<ID, Entity, Bean> extends BaseMapper<ID, Entity, Bean>{
}
public interface LifecycleBaseMapper<ID, Entity, Bean> extends BaseMapper<ID, Entity, Bean>{
}4、多数据源定义和动态切换
import org.springframework.jdbc.datasource.lookup.AbstractRoutingDataSource;
public class MultiDataSource extends AbstractRoutingDataSource {
public static void setDataSourceKey(String dataSource) {
MultiDataSourceHolder.setDataSource(dataSource);
}
@Override
protected Object determineCurrentLookupKey() {
return MultiDataSourceHolder.getDataSource();
}
}
public class MultiDataSourceHolder {
/**
* 注意：数据源标识保存在线程变量中，避免多线程操作数据源时互相干扰
*/
private static final ThreadLocal<String> dataSourceHolder = new ThreadLocal<String>();
public static String getDataSource() {
return dataSourceHolder.get();
}
public static void setDataSource(String dataSource) {
dataSourceHolder.set(dataSource);
}
public static void clearDataSource() {
dataSourceHolder.remove();
}
}
@Aspect
public class MultiDataSourceAspectAdvice {
private static final String JDBUY_DATA_SOURCE = "jdbuyDataSource";
private static final String LIFECYCLE_DATA_SOURCE = "lifecycleDataSource";
// @Pointcut("execution(* com.jd.zc.lifecycle.dao.*.*(..))")
// 切入点路径配置很关键,dao包下的所有子包、所有类、所有方法，".."所有参数
@Pointcut("execution(* com.jd.zc.lifecycle.dao.*.*.*(..))")
protected void pointcut() {
}
// @Around("execution(* com.jd.zc.lifecycle.*.*(..))")
@Around(value = "pointcut()")
public Object doAround(ProceedingJoinPoint jp) throws Throwable {
Object target = jp.getTarget();
if (target instanceof LifecycleBaseMapper) {
MultiDataSource.setDataSourceKey(LIFECYCLE_DATA_SOURCE);
} else if (target instanceof JdbuyBaseMapper) {
MultiDataSource.setDataSourceKey(JDBUY_DATA_SOURCE);
}
return jp.proceed();
}
}核心点就是，配置Mapper文件执行时的AOP，根据父类Mapper是LifecycleBaseMapper，还是JdbuyBaseMapper，动态切换数据源。
5、事务的定义
spring-info-datasource.xml已经定义了2个事务，实际使用的是时候，根据配置的标记来使用。
<qualifier value="lifecycle"/>
比如：
//需要指定使用的事务标记，2个数据源
@Transactional("lifecycle")
// 可以执行任务，就返回config配置
public boolean startTask(String key,Long timeout) {
logger.info(String.format("Try to get taskLock,key=%s,start",key));
// 简单锁1行，防止多个任务同时执行
TaskConfig config = configMapper.getForUpdate(key);
if (config == null) {
logger.error("The task config is null,key=" + key);
return false;
}
}
特别需要说的是，就算是配置了事务，它只是让属于1个数据库的数据源的CRUD在一个事务中，
多个事务没有归属到一个总体的事务中。
至于分布式事务，多数据源事务，还需要再研究。
---------------------
作者：小雷FansUnion
