java - 一级缓存不适用于 JPA 和 Hibernate

Question

我是一个使用 hibernate 缓存(第一级、第二级和查询缓存)的新手。

我的项目是使用 Spring MVC 和 JPA 配置的。

我正在使用下面的 JUnit 测试用例测试一级缓存。

public class FirstLevelCacheTest
{
    private static final Logger logger = LoggerFactory.getLogger(FirstLevelCacheTest.class);

    @PersistenceContext
    private EntityManager       entityManager;

    @Test
    public void testFirstLevelCacheTest()
    {
        Long clientId = 1L;

        // fetch the client entity from database first time
        Client client1 = entityManager.find(Client.class, clientId);
        logger.debug("Client Name : {}", client1.getName());

        // fetch the client entity again
        client1 = entityManager.find(Client.class, clientId);
        logger.debug("Client Name : {}", client1.getName());
    }
}

而我的实体类定义为:

@Entity
@JsonSerialize(include = JsonSerialize.Inclusion.NON_NULL)
@Table(name = "client")
public class Client extends BaseModel
{
    // instance variable declaration

    // getter and setter methods for instance variables
}

如果默认启用一级缓存，这应该执行一次 native 查询。 但我在执行此查询时得到以下结果:

Hibernate: select client0_.id as id0_0_, client0_.created as created0_0_, client0_.createdBy as createdBy0_0_, client0_.updated as updated0_0_, client0_.updatedBy as updatedBy0_0_, client0_.contactNo as contactNo0_0_, client0_.contactPerson as contactP7_0_0_, client0_.contactPersonEmail as contactP8_0_0_, client0_.contactPersonNo as contactP9_0_0_, client0_.email as email0_0_, client0_.name as name0_0_ from client client0_ where client0_.id=?
[main] DEBUG Client Name : Client1
Hibernate: select client0_.id as id0_0_, client0_.created as created0_0_, client0_.createdBy as createdBy0_0_, client0_.updated as updated0_0_, client0_.updatedBy as updatedBy0_0_, client0_.contactNo as contactNo0_0_, client0_.contactPerson as contactP7_0_0_, client0_.contactPersonEmail as contactP8_0_0_, client0_.contactPersonNo as contactP9_0_0_, client0_.email as email0_0_, client0_.name as name0_0_ from client client0_ where client0_.id=?
[main] DEBUG Client Name : Client1

下面是我的持久化相关配置:

@Configuration
@EnableTransactionManagement
public class PersistanceConfig
{
    // injection

    private String[] PACKAGES_TO_SCAN = new String[] { "com.mypackage1", "com.mypackage2" };

    @Bean
    public DataSource dataSource()
    {
        // dataSource setting 'com.mysql.jdbc.Driver' as a jdbc driver
    }

    private Properties jpaProperties()
    {
        Properties properties = new Properties();

        properties.put(AvailableSettings.DIALECT, hibernateDialect);
        properties.put(AvailableSettings.SHOW_SQL, hibernateShowSql);

        // 2nd level cache
        properties.put("hibernate.cache.use_second_level_cache", true);
        properties.put("hibernate.cache.region.factory_class", "org.hibernate.cache.ehcache.EhCacheRegionFactory");

        return properties;
    }

    @Bean
    public LocalContainerEntityManagerFactoryBean entityManagerFactory()
    {
        // entity manager settings using dataSource and jpaProperties
    }

    @Bean
    public JpaTransactionManager transactionManager()
    {
        // transaction manager settings using entityManagerFactory
    }
}

谁能帮我解决一个问题或我做错了什么？

提前致谢!

Answer 1

您需要使用 @Transactional 注释测试方法或使用 Spring TransactionTemplate .

即使检索实体不需要事务，将多个调用分组到一个内也会清楚地指定持久性上下文边界(它应该从哪里开始以及应该在哪里结束)。

要复制第一级缓存，您需要在整个测试方法中使用相同的工作单元，并且使用 @Transactional 注释测试将启动事务(绑定(bind)到当前执行的 Spring 逻辑事务与实际物理数据库事务相关的持久性上下文事务)。

当测试方法结束时，事务通常会回滚，因此不会将任何更改传播到任何后续测试，但在当前执行的测试方法中，您会看到自己的更改，即 ACID 中的 Isolation 属性。