SpringBoot实现动态数据源切换及单库事务控制
引言:
项目中经常会遇到多数据源的场景,通常的处理是: 操作数据一般都是在DAO层进行处理,使用配置多个DataSource 然后创建多个SessionFactory,在使用Dao层的时候通过不同的SessionFactory进行处理,
但是这样的操作代码入侵性比较明显且配置繁琐难以维护,,,在这里推荐一个Spring提供的AbstractRoutingDataSource抽象类,它实现了DataSource接口的用于获取数据库连接的方法
一. 注解方式实现动态切换数据源
原理: AbstractRoutingDataSource提供了程序运行时动态切换数据源的方法,在dao类或方法上标注需要访问数据源的关键字,路由到指定数据源,获取连接
1.pom.xml导入相关坐标
<!--mysql相关--> <dependency> <groupId>mysql</groupId> <artifactId>mysql-connector-java</artifactId> <scope>runtime</scope> </dependency> <!--oracle相关--> <dependency> <groupId>com.github.noraui</groupId> <artifactId>ojdbc7</artifactId> <version>${oracle.version}</version>
2.1application.properties配置多数据源
#多数据源配置 db.groups=default,oracle #默认数据库(mysql库) db.default.url=jdbc:mysql://127.0.0.1:3306/demo?connectTimeout=2000&allowMultiQueries=true&rewriteBatchedStatements=true&characterEncoding=UTF-8&serverTimezone=Asia/Shanghai db.default.username=root db.default.password=root #oracle库 db.oracle.url=jdbc:oracle:thin:@127.0.0.1:1521/orcl db.oracle.username=root db.oracle.password=root
2.2application.yml添加datasource配置
spring:
datasource:
group: ${db.groups}
3.1数据源切换方法: 维护一个static变量datasourceContext用于记录每个线程需要使用的数据源关键字。并提供切换、读取、清除数据源配置信息的方法
编写DataSourceContextHolder类
public class DataSourceContextHolder { private static final ThreadLocal<String> contextHolder = new ThreadLocal<String>(); public static synchronized void setDataSourceKey(String key) { contextHolder.set(key); } public static String getDataSourceKey() { return contextHolder.get(); } public static void clearDataSourceKey() { contextHolder.remove(); } }
3.2实现AbstractRoutingDataSource
public class DynamicDataSource extends AbstractRoutingDataSource {
private static DynamicDataSource instance; private static Object lock=new Object(); private static Map<Object,Object> dataSourceMap = Maps.newHashMap(); @Override protected Object determineCurrentLookupKey() { return DataSourceContextHolder.getDataSourceKey(); } @Override public void setTargetDataSources(Map<Object, Object> targetDataSources) { super.setTargetDataSources(targetDataSources); dataSourceMap.putAll(targetDataSources); super.afterPropertiesSet(); } public static synchronized DynamicDataSource getInstance(){ if(instance==null){ synchronized (lock){ if(instance==null){ instance=new DynamicDataSource(); } } } return instance; } public static boolean isExistDataSource(String key) { if (StringUtils.isEmpty(key)) { return false; } return dataSourceMap.containsKey(key); } }
3.3编写数据源配置类MybatisConfig
@Configuration @MapperScan(basePackages = { "com.**.mapper"} , sqlSessionTemplateRef = "sqlSessionTemplate")
public class MybatisConfig {
private static Logger LOG = LoggerFactory.getLogger(MybatisConfig.class);
@Autowired
private Environment environment;
private static final String DEFAULT_DATASOURCE_NAME = "default";
@Bean(name = "sqlSessionFactory")
public SqlSessionFactory sqlSessionFactory(@Qualifier("dynamicDataSource") DataSource dynamicDataSource) throws Exception {
//需要引入mybatis-plus坐标
//MybatisSqlSessionFactoryBean bean = new MybatisSqlSessionFactoryBean();
SqlSessionFactoryBean bean = new SqlSessionFactoryBean();
bean.setDataSource(dynamicDataSource);
//bean.setMapperLocations(new PathMatchingResourcePatternResolver().getResources("classpath*:com/wttech/vsm/support/mapper/*.xml"));
return bean.getObject();
}
@Bean(name = "transactionManager")
public PlatformTransactionManager transactionManager(@Qualifier("dynamicDataSource") DataSource dynamicDataSource) {
return new DataSourceTransactionManager(dynamicDataSource);
}
@Bean(name = "sqlSessionTemplate")
public SqlSessionTemplate sqlSessionTemplate(@Qualifier("sqlSessionFactory") SqlSessionFactory sqlSessionFactory) throws Exception {
return new SqlSessionTemplate(sqlSessionFactory);
}
@Bean(name = "dynamicDataSource")
public DynamicDataSource dynamicDataSource() {
String groups = environment.getProperty("spring.datasource.group");
LOG.info("数据源组名称:{}", groups);
Map<Object,Object> dataSourceMap = Maps.newHashMap();
Set<String> dbNames = Arrays.asList(groups.split(",")).stream().filter(s -> s.trim().length() > 0).collect(Collectors.toSet());
dbNames.add(DEFAULT_DATASOURCE_NAME);
HikariDataSource first = null;
HikariDataSource def = null;
for (String dbName:dbNames) {
String driver = environment.getProperty(String.format("db.%s.driver", dbName));
String url = environment.getProperty(String.format("db.%s.url", dbName));
String username = environment.getProperty(String.format("db.%s.username", dbName));
String password = environment.getProperty(String.format("db.%s.password", dbName));
LOG.info("数据源{}连接:{}", dbName, url);
if (StringUtils.isEmpty(url) || StringUtils.isEmpty(username) || StringUtils.isEmpty(password)) {
continue;
}
DataSourceBuilder<HikariDataSource> hikariDataSourceBuilder = DataSourceBuilder.create().type(HikariDataSource.class);
if (!StringUtils.isEmpty(driver)) {
hikariDataSourceBuilder.driverClassName(driver);
}
HikariDataSource hikariDataSource = hikariDataSourceBuilder.url(url).username(username).password(password).build();
hikariDataSource.setAutoCommit(true);
String testQuery = environment.getProperty(String.format("db.%s.connectionTestQuery", dbName));
if (!StringUtils.isEmpty(testQuery)) {
hikariDataSource.setConnectionTestQuery(testQuery);
}
String timeout = environment.getProperty(String.format("db.%s.connectionTimeout", dbName));
if (!StringUtils.isEmpty(timeout)) {
hikariDataSource.setConnectionTimeout(Long.parseLong(timeout));
}
String minimumIdle = environment.getProperty(String.format("db.%s.minimumIdle", dbName));
if (!StringUtils.isEmpty(minimumIdle)) {
hikariDataSource.setMinimumIdle(Integer.parseInt(minimumIdle));
}
String maximumPoolSize = environment.getProperty(String.format("db.%s.maximumPoolSize", dbName));
if (!StringUtils.isEmpty(maximumPoolSize)) {
hikariDataSource.setMaximumPoolSize(Integer.parseInt(maximumPoolSize));
}
String idleTimeout = environment.getProperty(String.format("db.%s.idleTimeout", dbName));
if (!StringUtils.isEmpty(idleTimeout)) {
hikariDataSource.setIdleTimeout(Long.parseLong(idleTimeout));
}
String maxLifetime = environment.getProperty(String.format("db.%s.maxLifetime", dbName));
if (!StringUtils.isEmpty(maxLifetime)) {
hikariDataSource.setMaxLifetime(Long.parseLong(maxLifetime));
}
hikariDataSource.setPoolName(dbName);
dataSourceMap.put(dbName, hikariDataSource);
if (first == null) {
first = hikariDataSource;
}
if (DEFAULT_DATASOURCE_NAME.equals(dbName)) {
def = hikariDataSource;
}
}
DynamicDataSource dynamicDataSource = DynamicDataSource.getInstance();
dynamicDataSource.setTargetDataSources(dataSourceMap);
dynamicDataSource.setDefaultTargetDataSource(def == null ? first : def);
return dynamicDataSource;
}
4.1.标记数据源注解
@Target({ElementType.METHOD, ElementType.TYPE}) @Retention(RetentionPolicy.RUNTIME) public @interface SwitchDataSource { String value(); }
4.2.编写切入点方法
@Aspect @Component public class MethodInterceptor { @Around("execution(* com.wttech.vsm.support.mapper..*.*(..))) public Object dao(ProceedingJoinPoint invocation) throws Throwable { MethodSignature methodSignature = (MethodSignature) invocation.getSignature(); Method method = methodSignature.getMethod(); String dbName = null; SwitchDataSource dataSource = method.getAnnotation(SwitchDataSource.class); if (dataSource != null) { dbName = dataSource.value(); if (DynamicDataSource.isExistDataSource(dbName)) { DataSourceContextHolder.setDataSourceKey(dbName); } } try { return invocation.proceed(); } finally { DataSourceContextHolder.clearDataSourceKey(); } } }
二.单库事务控制:
本人涉及到到的业务场景是在service层的一个方法中存在多个dao操作(只涉及单库),,需要维持事务性,,,遇到问题: 数据源是在mapper层通过注解切换的,,@Transactional在services层控制,,导致程序报错找不到数据源,,开始的解决思路是前提@SwitchDataSource注解至service层 ,,但经测试后还是报错,,,
最后的解决思路是必须保证切换数据源是在事务控制开启之前完成...
1.切入点表达式增加service层切入
@Around("execution(* com.**.mapper..*.*(..)) || execution(* com.**.service.DataFixService.*(..))")
2.注释掉需要事务控制的dao操作设计到的mapper层的数据源切换注解
// @SwitchDataSource("toll") int updateFixInList(@Param("listId") String listId, @Param("fieldName") String fieldName, @Param("fieldValue") String fieldValue);
3.事务控制是基于数据源的,,必须在数据源切换后在开启事务,,以下是具体实现思路:
控制器->方法1(切换数据源,使用代理方式调用方法2)->方法2(开启事务,执行多个dao操作)
先切换数据源
@SwitchDataSource("toll") public void updateFixInList(String listId, String fieldName, String fieldValue) { //springAOP的用法中,只有代理的类才会被切入,我们在controller层调用service的方法的时候,是可以被切入的,但是如果我们在service层 A方法中,调用B方法,切点切的是B方法,那么这时候是不会切入的 //通过((Service)AopContext.currentProxy()).B() 来调用B方法,这样一来,就能切入了! ((DataFixService) AopContext.currentProxy()).updateFixInListProxy(listId, fieldName, fieldValue); }
再控制事务
@Transactional(rollbackFor = Exception.class) public void updateFixInListProxy(String listId, String fieldName, String fieldValue) { dataFixMapper.updateFixInList(listId, fieldName, fieldValue); //其他dao操作 }
特别注意: 以上这种方法目前仅适用于多数据源下单库的事务操作,,,如果serviece方法dao操作设计多库,,由于目前业务场景暂未涉及到,,所以暂未深入研究.....
【推荐】国内首个AI IDE,深度理解中文开发场景,立即下载体验Trae
【推荐】编程新体验,更懂你的AI,立即体验豆包MarsCode编程助手
【推荐】抖音旗下AI助手豆包,你的智能百科全书,全免费不限次数
【推荐】轻量又高性能的 SSH 工具 IShell:AI 加持,快人一步
· Linux系列:如何用heaptrack跟踪.NET程序的非托管内存泄露
· 开发者必知的日志记录最佳实践
· SQL Server 2025 AI相关能力初探
· Linux系列:如何用 C#调用 C方法造成内存泄露
· AI与.NET技术实操系列(二):开始使用ML.NET
· 无需6万激活码!GitHub神秘组织3小时极速复刻Manus,手把手教你使用OpenManus搭建本
· C#/.NET/.NET Core优秀项目和框架2025年2月简报
· Manus爆火,是硬核还是营销?
· 一文读懂知识蒸馏
· 终于写完轮子一部分:tcp代理 了,记录一下