Spring多数据源实现的方式大概有2中,一种是新建多个 MapperScan
扫描不同包,另外一种则是通过继承 AbstractRoutingDataSource
实现动态路由。今天作者主要基于后者做的实现,且方式1的实现比较简单这里不做过多探讨。
@Configuration @MapperScan(basePackages = "com.goofly.test1", sqlSessionTemplateRef = "test1SqlSessionTemplate") public class DataSource1Config1 { @Bean(name = "dataSource1") @ConfigurationProperties(prefix = "spring.datasource.test1") @Primary public DataSource testDataSource() { return DataSourceBuilder.create().build(); } // .....略 } @Configuration @MapperScan(basePackages = "com.goofly.test2", sqlSessionTemplateRef = "test1SqlSessionTemplate") public class DataSourceConfig2 { @Bean(name = "dataSource2") @ConfigurationProperties(prefix = "spring.datasource.test2") @Primary public DataSource testDataSource() { return DataSourceBuilder.create().build(); } // .....略 }
public class DynamicRoutingDataSource extends AbstractRoutingDataSource { private static final Logger log = Logger.getLogger(DynamicRoutingDataSource.class); @Override protected Object determineCurrentLookupKey() { //从ThreadLocal中取值 return DynamicDataSourceContextHolder.get(); } }
第1种方式虽然实现比较加单,劣势就是不同数据源的mapper文件不能在同一包名,就显得不太灵活了。所以为了更加灵活的作为一个组件的存在,作者采用的第二种方式实现。
ThreadLocal
中 AbstractRoutingDataSource.determineCurrentLookupKey()
方法,然后从 ThreadLocal
中获取之前设置的key值,然后将该值返回。 AbstractRoutingDataSource
的 targetDataSources
是一个map,保存了数据源key和数据源的对应关系,所以能够顺利的找到该对应的数据源。 org.springframework.jdbc.datasource.lookup.AbstractRoutingDataSource
,如下:
public abstract class AbstractRoutingDataSource extends AbstractDataSource implements InitializingBean { private Map<Object, Object> targetDataSources; private Object defaultTargetDataSource; private boolean lenientFallback = true; private DataSourceLookup dataSourceLookup = new JndiDataSourceLookup(); private Map<Object, DataSource> resolvedDataSources; private DataSource resolvedDefaultDataSource; protected DataSource determineTargetDataSource() { Assert.notNull(this.resolvedDataSources, "DataSource router not initialized"); Object lookupKey = determineCurrentLookupKey(); DataSource dataSource = this.resolvedDataSources.get(lookupKey); if (dataSource == null && (this.lenientFallback || lookupKey == null)) { dataSource = this.resolvedDefaultDataSource; } if (dataSource == null) { throw new IllegalStateException("Cannot determine target DataSource for lookup key [" + lookupKey + "]"); } return dataSource; } /** * Determine the current lookup key. This will typically be * implemented to check a thread-bound transaction context. * <p>Allows for arbitrary keys. The returned key needs * to match the stored lookup key type, as resolved by the * {@link #resolveSpecifiedLookupKey} method. */ protected abstract Object determineCurrentLookupKey(); //........略
targetDataSources
是一个map结构,保存了key与数据源的对应关系;
dataSourceLookup
是一个 DataSourceLookup
类型,默认实现是 JndiDataSourceLookup
。点开该类源码会发现,它实现了通过key获取DataSource的逻辑。当然,这里可以通过 setDataSourceLookup()
来改变其属性,因为关于此处有一个坑,后面会讲到。
public class JndiDataSourceLookup extends JndiLocatorSupport implements DataSourceLookup { public JndiDataSourceLookup() { setResourceRef(true); } @Override public DataSource getDataSource(String dataSourceName) throws DataSourceLookupFailureException { try { return lookup(dataSourceName, DataSource.class); } catch (NamingException ex) { throw new DataSourceLookupFailureException( "Failed to look up JNDI DataSource with name '" + dataSourceName + "'", ex); } } }
# db1 spring.datasource.master.url = jdbc:mysql://127.0.0.1:3306/test?useUnicode=true&characterEncoding=utf8&useSSL=false spring.datasource.master.username = root spring.datasource.master.password = 123456 spring.datasource.master.driverClassName = com.mysql.jdbc.Driver spring.datasource.master.validationQuery = true spring.datasource.master.testOnBorrow = true ## db2 spring.datasource.slave.url = jdbc:mysql://127.0.0.1:3306/test1?useUnicode=true&characterEncoding=utf8&useSSL=false spring.datasource.slave.username = root spring.datasource.slave.password = 123456 spring.datasource.slave.driverClassName = com.mysql.jdbc.Driver spring.datasource.slave.validationQuery = true spring.datasource.slave.testOnBorrow = true #主数据源名称 spring.maindb=master #mapperper包路径 mapper.basePackages =com.btps.xli.multidb.demo.mapper
### 单数据源 为了让使用者能够用最小的改动实现最好的效果,作者对单数据源的多种配置做了兼容。 配置1:
spring.datasource.master.url = jdbc:mysql://127.0.0.1:3306/test?useUnicode=true&characterEncoding=utf8&useSSL=false
spring.datasource.master.username = root
spring.datasource.master.password = 123456
spring.datasource.master.driverClassName = com.mysql.jdbc.Driver
spring.datasource.master.validationQuery = true
spring.datasource.master.testOnBorrow = true
mapper.basePackages = com.goofly.xli.multidb.demo.mapper
spring.maindb=master
配置2
spring.datasource.url = jdbc:mysql://127.0.0.1:3306/test?useUnicode=true&characterEncoding=utf8&useSSL=false
spring.datasource.username = root
spring.datasource.password = 123456
spring.datasource.driverClassName = com.mysql.jdbc.Driver
spring.datasource.validationQuery = true
spring.datasource.testOnBorrow = true
mapper.basePackages = com.goofly.xli.multidb.demo.mapper
## 踩坑之路 ### 多数据源的循环依赖
Description:
The dependencies of some of the beans in the application context form a cycle:
happinessController (field private com.db.service.HappinessService com.db.controller.HappinessController.happinessService)
↓
happinessServiceImpl (field private com.db.mapper.MasterDao com.db.service.HappinessServiceImpl.masterDao)
↓
masterDao defined in file [E:GitRepositoryframework-graytest-dbtargetclassescomdbmapperMasterDao.class]
↓
sqlSessionFactory defined in class path resource [com/goofly/xli/datasource/core/DynamicDataSourceConfiguration.class]
┌─────┐
| dynamicDataSource defined in class path resource [com/goofly/xli/datasource/core/DynamicDataSourceConfiguration.class]
↑ ↓
| firstDataSource defined in class path resource [com/goofly/xli/datasource/core/DynamicDataSourceConfiguration.class]
↑ ↓
| dataSourceInitializer
**解决方案:** 在Spring boot启动的时候排除`DataSourceAutoConfiguration`即可。如下:
@SpringBootApplication(exclude = {DataSourceAutoConfiguration.class})
public class DBMain {
public static void main(String[] args) { SpringApplication.run(DBMain.class, args); }
}
> 但是作者在创建多数据源的时候由于并未创建多个`DataSource`的Bean,而是只创建了一个即需要做动态数据源的那个Bean。 其他的`DataSource`则直接创建实例然后存放在Map里面,然后再设置到`DynamicRoutingDataSource#setTargetDataSources`即可。 > > 因此这种方式也不会出现循环依赖的问题! ### 动态刷新数据源 > 笔者在设计之初是想构建一个动态刷新数据源的方案,所以利用了`SpringCloud`的`@RefreshScope`去标注数据源,然后利用`RefreshScope#refresh`实现刷新。但是在实验的时候发现由Druid创建的数据源会因此而关闭,由Spring的`DataSourceBuilder`创建的数据源则不会发生任何变化。 > > 最后关于此也没能找到解决方案。同时思考,如果只能的可以实现动态刷新的话,那么数据源的原有连接会因为刷新而中断吗还是会有其他处理? ### 多数据源事务 > 有这么一种特殊情况,一个事务中调用了两个不同数据源,这个时候动态切换数据源会因此而失效。 > > 翻阅了很多文章,大概找了2中解决方案,一种是Atomikos进行事务管理,但是貌似性能并不是很理想。 > > 另外一种则是通过优先级控制,切面的的优先级必须要大于数据源的优先级,用注解`@Order`控制。 >