一、背景以及为什么需要学习
在高并发的项目中,单数据库已无法承载大数据量的访问,因此需要使用多个数据库进行对数据的读写分离,此外就是在微服化的今天,我们在项目中可能采用各种不同存储,因此也需要连接不同的数据库,居于这样的背景,这里简单分享实现的思路以及实现方案
二、实现方式
多数据源实现思路有两种,一种是通过配置多个SqlSessionFactory实现多数据源;
1、通过配置多个SqlSessionFactory实现多数据源
2、通过Spring提供的AbstractRoutingDataSource抽象了一个DynamicDataSource实现动态切换数据源
三、方式一:不同库的Mapper指定不同的SqlSessionFactory
1 针对不同的库分别放置对用不同的SqlSessionFactory
UserDataSourceConfiguration
@Configuration
@MapperScan(basePackages = "org.datasource.demo1.usermapper",
sqlSessionFactoryRef = "userSqlSessionFactory")
public class UserDataSourceConfiguration {
public static final String MAPPER_LOCATION = "classpath:usermapper/*.xml";
@Primary
@Bean("userDataSource")
@ConfigurationProperties(prefix = "spring.datasource.user")
public DataSource userDataSource() {
return DataSourceBuilder.create().build();
}
@Bean(name = "userTransactionManager")
@Primary
public PlatformTransactionManager userTransactionManager(@Qualifier("userDataSource") DataSource dataSource) {
return new DataSourceTransactionManager(dataSource);
}
@Primary
@Bean(name = "userSqlSessionFactory")
public SqlSessionFactory userSqlSessionFactory(@Qualifier("userDataSource") DataSource dataSource) throws Exception {
final SqlSessionFactoryBean sessionFactoryBean = new SqlSessionFactoryBean();
sessionFactoryBean.setDataSource(dataSource);
sessionFactoryBean.setMapperLocations(new PathMatchingResourcePatternResolver().getResources(UserDataSourceConfiguration.MAPPER_LOCATION));
return sessionFactoryBean.getObject();
}
}
SoulDataSourceConfiguration
@Configuration
@MapperScan(basePackages = "org.datasource.demo1.soulmapper",
sqlSessionFactoryRef = "soulSqlSessionFactory")
public class SoulDataSourceConfiguration {
public static final String MAPPER_LOCATION = "classpath:soulmapper/*.xml";
@Bean("soulDataSource")
@ConfigurationProperties(prefix = "spring.datasource.soul")
public DataSource soulDataSource() {
return DataSourceBuilder.create().build();
}
@Bean(name = "soulTransactionManager")
public PlatformTransactionManager soulTransactionManager(@Qualifier("soulDataSource") DataSource dataSource) {
return new DataSourceTransactionManager(dataSource);
}
@Bean(name = "soulSqlSessionFactory")
public SqlSessionFactory soulSqlSessionFactory(@Qualifier("soulDataSource") DataSource dataSource) throws Exception {
final SqlSessionFactoryBean sessionFactoryBean = new SqlSessionFactoryBean();
sessionFactoryBean.setDataSource(dataSource);
sessionFactoryBean.setMapperLocations(new PathMatchingResourcePatternResolver().getResources(SoulDataSourceConfiguration.MAPPER_LOCATION));
return sessionFactoryBean.getObject();
}
}
2 使用和测试
@Service
public class AppAuthService {
@Autowired
private AppAuthMapper appAuthMapper;
@Transactional(rollbackFor = Exception.class)
public int getCount() {
int a = appAuthMapper.listCount();
int b = 1 / 0;
return a;
}
}
@SpringBootTest
@RunWith(SpringRunner.class)
public class TestDataSource {
@Autowired
private AppAuthService appAuthService;
@Autowired
private SysUserService sysUserService;
@Test
public void test_dataSource1(){
int b=sysUserService.getCount();
int a=appAuthService.getCount();
}
}
3 总结
此种方式使用起来分层明确,不存在任何冗余代码,不足地方就是每个库都需要对应一个配置类,该配置类中实现方式都基本类似,该种解决方案每个配置类中都存在事务管理器,因此不需要单独再去额外的关注。在使用时需要指定事务管理器
四、AOP+自定义注解
关于采用Spring AOP方式实现原理就是把多个数据源存储在一个 Map中,当需要使用某个数据源时,从 Map中获取此数据源进行处理。
1 AbstractRoutingDataSource
在Spring中提供了AbstractRoutingDataSource来实现此功能,继承AbstractRoutingDataSource类并覆写其determineCurrentLookupKey()方法就可以完成数据源切换,该方法只需要返回数据源key即可,也就是存放数据源的Map的key,接下来我们来看一下AbstractRoutingDataSource整体的继承结构,看他是如何做到的。
在整体的继承结构上我们会发现AbstractRoutingDataSource最终是继承于DataSource,因此当我们继承AbstractRoutingDataSource是我们自身也是一个数据源,对于数据源必然有连接数据库的动作,如下代码:
public Connection getConnection() throws SQLException {
return this.determineTargetDataSource().getConnection();
}
public Connection getConnection(String username, String password) throws SQLException {
return this.determineTargetDataSource().getConnection(username, password);
}
只是AbstractRoutingDataSource的getConnection()方法里实际是调用determineTargetDataSource()返回的数据源的getConnection()方法。
protected DataSource determineTargetDataSource() {
Assert.notNull(this.resolvedDataSources, "DataSource router not initialized");
Object lookupKey = this.determineCurrentLookupKey();
DataSource 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 + "]");
} else {
return dataSource;
}
}
该方法通过determineCurrentLookupKey()方法获取一个key,通过key从resolvedDataSources中获取数据源DataSource对象。determineCurrentLookupKey()是个抽象方法,需要继承AbstractRoutingDataSource的类实现;而resolvedDataSources是一个Map<Object, DataSource>,里面应该保存当前所有可切换的数据源。
2 以下是项目中用到的目录结构
3 DataSourceType
public enum DataSourceType {
LOCALDATASOURCE("localDataSource"),
APPDATASOURCE("appDataSource"),
BASEDATASOURCE("baseDataSource"),
REMOTEDATASOURCE("remoteDataSource"),
WXTJDATASOURCE("wxtjDataSource"),
;
private String name;
DataSourceType(String name) {
this.name=name;
}
public String getName() {
return name;
}
public void setName(String name) {
this.name = name;
}
}
4 DynamicDataSourceConfiguration
通过读取配置文件中的数据源配置信息,创建数据连接,将多个数据源放入Map中,注入到容器中:
@Configuration
@MapperScan(basePackages = {
"org.datasource.demo.mapper"})
public class DynamicDataSourceConfiguration {
@Bean(name = "local")
@ConfigurationProperties(prefix = "local.datasource")
public DataSource userDataSource() {
DruidDataSource dataSource = DruidDataSourceBuilder.create().build();
return dataSource;
}
@Primary
@Bean(name = "app")
@ConfigurationProperties(prefix = "app-local.datasource")
public DataSource appLocalDataSource() {
DruidDataSource dataSource = DruidDataSourceBuilder.create().build();
return dataSource;
}
@Bean(name = "base")
@ConfigurationProperties(prefix = "base.datasource")
public DataSource appDataSource() {
DruidDataSource dataSource = DruidDataSourceBuilder.create().build();
return dataSource;
}
@Bean(name = "remote")
@ConfigurationProperties(prefix = "remote.datasource")
public DataSource remoteDataSource() {
DruidDataSource dataSource = DruidDataSourceBuilder.create().