Hibernate ehcache缓存配置详解+实战

Posted 2022-12-03 滕哥

依赖管理

我这里使用的maven来管理的依赖，pom如下：

<dependency>
    <groupId>org.hibernate</groupId>
    <artifactId>hibernate-core</artifactId>
    <version>5.0.2.Final</version>
</dependency>
<dependency>
    <groupId>org.hibernate</groupId>
    <artifactId>hibernate-ehcache</artifactId>
    <version>5.0.2.Final</version>
</dependency>

开始配置缓存

使用ehcache需要配置文件：ehcache.xml 并修改hibernate的配置

ehcache.xml配置详解

配置文件如下：

<ehcache>　　
　　<!-- 
　　指定二级缓存存放在磁盘上的位置，可以使用磁盘目录，也可以使用Java System Property目录，user.home是用户目录、user.dir是用户当前工作目录、java.io.tmpdir是默认临时文件路径 
　　-->
   <diskStore path="java.io.tmpdir/cache"/>

　　<defaultCache
        maxElementsInMemory="10000"
        eternal="false"
        timeToIdleSeconds="120"
        timeToLiveSeconds="120"
        overflowToDisk="true"
        />

　　<!--可以给每个实体类指定一个配置文件，通过name属性指定，要使用类的全名-->
    <cache name="com.teemo.entity.User"
        maxElementsInMemory="10000"
        eternal="false"
        timeToIdleSeconds="300"
        timeToLiveSeconds="600"
        overflowToDisk="true"
        />

    <cache name="sample"
        maxElementsInMemory="1000"
        eternal="true"
        timeToIdleSeconds="0"
        timeToLiveSeconds="0"
        overflowToDisk="false"
        />
</ehcache>

cache字段说明：

1. name：Cache的唯一标识。
2. maxElementsInMemory：内存中最大缓存对象数。
3. eternal：Element是否永久有效，一旦设置true，timeout将不起作用。
4. timeToIdleSeconds：设置Element在失效前的允许闲置时间。仅当element不是永久有效时使用，可选属性，默认值是0，也就是可闲置时间无穷大。
5. timeToLiveSeconds：设置Element在失效前允许存活时间。最大时间介于创建时间和失效时间之间。仅当element不是永久有效时使用，默认是0.，也就是element存活时间无穷大。
6. overflowToDisk：配置此属性，当内存中Element数量达到maxElementsInMemory时，Ehcache将会Element写到磁盘中。
7. maxElementsOnDisk：磁盘中最大缓存对象数，若是0表示无穷大。
8. memoryStoreEvictionPolicy：当达到maxElementsInMemory限制时，Ehcache将会根据指定的策略
9. 去清理缓存中的内容。默认策略是LRU（最近最少使用），你也可以设置为FIFO（先进先出）或是LFU（较少使用）

Hibernate二级缓存逻辑：

会先根据entity的region到ehcache的缓存区域查找缓存，region对应的是ehcache配置中的name字段，如果找不到，会使用entity的全类名来查找缓存，如果还找不到会使用default设置

Hibernate缓存设置

xml配置如下：

<!-- 开启二级缓存 -->
<prop key="hibernate.cache.use_second_level_cache">true</prop>
<!-- 开启查询缓存 -->
<prop key="hibernate.cache.use_query_cache">true</prop>
<!-- 高速缓存提供程序 -->
<!-- Hibernate4.0 以前使用该设置<prop key="hibernate.cache.provider_class">net.sf.ehcache.hibernate.SingletonEhCacheProvider</prop>-->
<prop key="hibernate.cache.region.factory_class">org.hibernate.cache.ehcache.SingletonEhCacheRegionFactory</prop>

实体类注解：

@Cache(region = "sample", usage = CacheConcurrencyStrategy.READ_WRITE)
public class User 
    @Cache(region = "all", usage = CacheConcurrencyStrategy.READ_WRITE)
    private Set<Role> roles = new HashSet<Role>();

usage提供的事务隔离级别有NONE, READ_ONLY, NONSTRICT_READ_WRITE, READ_WRITE, TRANSACTIONAL，但是ehcache不支持TRANSACTIONAL。

使用并测试缓存效果

二级缓存适用于以下数据：

1. 很少被修改，大量查询的
2. 不是很重要的数据，允许出现偶尔并发访问的

Hibernate的缓存模式

Hibernate的二级缓存只会根据ID进行缓存，也就是使用get/load方法会将查询对象缓存，如果需要对HQL或SQL进行缓存，则需要在代码中指定setCacheable(true)。
查询缓存所缓存的key值就是查询所使用的HQL或SQL语句，需要注意的是：查询缓存不仅要求所使用的HQL语句、SQL语句相同，甚至是要求所传入的参数也相同，Hibernate才能从缓存中查去数据。

1、get方法缓存测试

测试代码：

    /**
     * service.get方法直接使用Hibernate Session.get(class, id)
     */
    @Test
    public void testGet() 
        DynamicProperty dp1 = service.get(1L);
        System.out.println(dp1);
        DynamicProperty dp2 = service.get(1L);
        System.out.println(dp2);
    

输出结果：

Hibernate: select dynamicpro0_.id as id1_1_0_, dynamicpro0_.author as author2_1_0_（省略）from dynamic_property dynamicpro0_ where dynamicpro0_.id=?
com.teemo.entity.DynamicProperty@5147788d
com.teemo.entity.DynamicProperty@50ab025b

从结果中可以看到，第二次查询命中缓存，没有发出SQL查询。

2、根据字段查询，不指定setCacheable(true)

测试代码：

    /**
     * service.get方法使用Query query = getSession().createQuery(hql);
     * query.list来返回结果，没有指定query的setCacheable
     */
    @Test
    public void testGetByProperty() 
        String key = "System.Version";
        DynamicProperty dp1 = service.get("dynamicPropertyKey", key);
        System.out.println(dp1);
        DynamicProperty dp2 = service.get("dynamicPropertyKey", key);
        System.out.println(dp2);
    

输出结果：

Hibernate: select dynamicpro0_.id as id1_1_, dynamicpro0_.author as author2_1_（省略）from dynamic_property dynamicpro0_ where dynamicpro0_.property_key=?
com.teemo.entity.DynamicProperty@5e7e7944

Hibernate: select dynamicpro0_.id as id1_1_, dynamicpro0_.author as author2_1_（省略）from dynamic_property dynamicpro0_ where dynamicpro0_.property_key=?
com.teemo.entity.DynamicProperty@74542394

对于没有指定query.setCacheable(true)的查询，没有使用缓存对象

3、根据字段查询，不指定setCacheable(true)，但是主对象包含List/Set

List/Set中的对象，属于查询缓存

测试代码：

    @Test
    public void testGetByEmail() 
        String email = "[e-mail]";
        User user1 = userService.get("email", email);
        System.out.println(user1);
        User user2 = userService.get("email", email);
        System.out.println(user2);
    

输出结果：

Hibernate: select user0_.id as id1_6_, user0_.create_time as create_t2_6_（省略）from user user0_ where user0_.email=?

Hibernate: select roles0_.user_id as user_id1_8_0_, roles0_.role_id as role_id2_8_0_（省略）from user_role roles0_ inner join role role1_ on roles0_.role_id=role1_.id where roles0_.user_id=?

Hibernate: select resourcepe0_.role_id as role_id4_5_0_, resourcepe0_.id as id1_5_0_, resourcepe0_.id as id1_5_1_（省略）from role_resource_permission resourcepe0_ where resourcepe0_.role_id=?
com.teemo.entity.User@1e6923ba

Hibernate: select user0_.id as id1_6_, user0_.create_time as create_t2_6_（省略）from user user0_ where user0_.email=?
com.teemo.entity.User@4cd29ffa

从结果中可以看出，在第一次查询User实体时，会出现1+N的情况，先查出User实体，在根据User实体查出关联Role，但是此时会把Role的id放到查询缓存中，对象放到二级缓存中，在第二次查询时，发现查询缓存存在这些Role id，那么就会到二级缓存中查询到响应的对象，所以第二次查询没有发出Role相关的SQL。

4、根据字段查询，指定setCacheable(true)

测试代码：

    @Test
    public void testGetByEmail() 
        String email = "[e-mail]";
        User user1 = userService.get("email", email);
        System.out.println(user1);
        User user2 = userService.get("email", email);
        System.out.println(user2);
    

输出结果：

2017-01-10 16:21:19.843 [main] DEBUG org.hibernate.cache.internal.StandardQueryCache  - Checking cached query results in region: org.hibernate.cache.internal.StandardQueryCache
2017-01-10 16:21:19.843 [main] DEBUG org.hibernate.cache.ehcache.internal.regions.EhcacheGeneralDataRegion  - key: sql: select user0_.id as id1_6_, user0_.create_time as create_t2_6_, user0_.deleted as deleted3_6_, user0_.department_key as departme4_6_, user0_.email as email5_6_, user0_.mobile_phone as mobile_p6_6_, user0_.modify_time as modify_t7_6_, user0_.nickname as nickname8_6_, user0_.password as password9_6_, user0_.salt as salt10_6_, user0_.status as status11_6_, user0_.username as usernam12_6_ from user user0_ where user0_.email=?; parameters: ; named parameters: email_0=[e-mail]; transformer: org.hibernate.transform.CacheableResultTransformer@110f2
2017-01-10 16:21:19.844 [main] DEBUG org.hibernate.cache.ehcache.internal.regions.EhcacheGeneralDataRegion  - Element for key sql: select user0_.id as id1_6_, user0_.create_time as create_t2_6_, user0_.deleted as deleted3_6_, user0_.department_key as departme4_6_, user0_.email as email5_6_, user0_.mobile_phone as mobile_p6_6_, user0_.modify_time as modify_t7_6_, user0_.nickname as nickname8_6_, user0_.password as password9_6_, user0_.salt as salt10_6_, user0_.status as status11_6_, user0_.username as usernam12_6_ from user user0_ where user0_.email=?; parameters: ; named parameters: email_0=[e-mail]; transformer: org.hibernate.transform.CacheableResultTransformer@110f2 is null
2017-01-10 16:21:19.845 [main] DEBUG org.hibernate.cache.internal.StandardQueryCache  - Query results were not found in cache
Hibernate: select user0_.id as id1_6_, user0_.create_time as create_t2_6_, user0_.deleted as deleted3_6_, user0_.department_key as departme4_6_, user0_.email as email5_6_, user0_.mobile_phone as mobile_p6_6_, user0_.modify_time as modify_t7_6_, user0_.nickname as nickname8_6_, user0_.password as password9_6_, user0_.salt as salt10_6_, user0_.status as status11_6_, user0_.username as usernam12_6_ from user user0_ where user0_.email=?
Hibernate: select roles0_.user_id as user_id1_8_0_, roles0_.role_id as role_id2_8_0_, role1_.id as id1_4_1_, role1_.available as availabl2_4_1_, role1_.description as descript3_4_1_, role1_.role_key as role_key4_4_1_, role1_.role_value as role_val5_4_1_ from user_role roles0_ inner join role role1_ on roles0_.role_id=role1_.id where roles0_.user_id=?
Hibernate: select resourcepe0_.role_id as role_id4_5_0_, resourcepe0_.id as id1_5_0_, resourcepe0_.id as id1_5_1_, resourcepe0_.permission_ids as permissi2_5_1_, resourcepe0_.resource_id as resource3_5_1_, resourcepe0_.role_id as role_id4_5_1_ from role_resource_permission resourcepe0_ where resourcepe0_.role_id=?
2017-01-10 16:21:20.064 [main] DEBUG org.hibernate.cache.internal.StandardQueryCache  - Caching query results in region: org.hibernate.cache.internal.StandardQueryCache; timestamp=6078613420920832
2017-01-10 16:21:20.064 [main] DEBUG org.hibernate.cache.ehcache.internal.regions.EhcacheGeneralDataRegion  - key: sql: select user0_.id as id1_6_, user0_.create_time as create_t2_6_, user0_.deleted as deleted3_6_, user0_.department_key as departme4_6_, user0_.email as email5_6_, user0_.mobile_phone as mobile_p6_6_, user0_.modify_time as modify_t7_6_, user0_.nickname as nickname8_6_, user0_.password as password9_6_, user0_.salt as salt10_6_, user0_.status as status11_6_, user0_.username as usernam12_6_ from user user0_ where user0_.email=?; parameters: ; named parameters: email_0=[e-mail]; transformer: org.hibernate.transform.CacheableResultTransformer@110f2 value: [6078613420920832, 1]

// 查询出第一个对象，并缓存
com.teemo.entity.User@1a115fb4

2017-01-10 16:21:20.077 [main] DEBUG org.hibernate.cache.internal.StandardQueryCache  - Checking cached query results in region: org.hibernate.cache.internal.StandardQueryCache
2017-01-10 16:21:20.077 [main] DEBUG org.hibernate.cache.ehcache.internal.regions.EhcacheGeneralDataRegion  - key: sql: select user0_.id as id1_6_, user0_.create_time as create_t2_6_, user0_.deleted as deleted3_6_, user0_.department_key as departme4_6_, user0_.email as email5_6_, user0_.mobile_phone as mobile_p6_6_, user0_.modify_time as modify_t7_6_, user0_.nickname as nickname8_6_, user0_.password as password9_6_, user0_.salt as salt10_6_, user0_.status as status11_6_, user0_.username as usernam12_6_ from user user0_ where user0_.email=?; parameters: ; named parameters: email_0=[e-mail]; transformer: org.hibernate.transform.CacheableResultTransformer@110f2
2017-01-10 16:21:20.078 [main] DEBUG org.hibernate.cache.internal.StandardQueryCache  - Checking query spaces are up-to-date: [user]
2017-01-10 16:21:20.079 [main] DEBUG org.hibernate.cache.ehcache.internal.regions.EhcacheGeneralDataRegion  - key: user
2017-01-10 16:21:20.079 [main] DEBUG org.hibernate.cache.ehcache.internal.regions.EhcacheGeneralDataRegion  - Element for key user is null
2017-01-10 16:21:20.080 [main] DEBUG org.hibernate.cache.internal.StandardQueryCache  - Returning cached query results

// 根据SQL/HQL比对缓存key，命中
com.teemo.entity.User@5b867c31

从结果中看出，在query中设置query.setCacheable(true);之后，所有被标记为可缓存的对象都会被缓存下来，所以第二次查询User时，直接命中二级缓存，没有发出SQL查询。

小结

我觉得，Hibernate的缓存模式应用场景太单一，对于通用的DAO或者底层查询方法来说，如果设置为query.setCacheable(true) 显然不太灵活，这样会缓存所有的调用该方法查询到的数据，如果不设置的话，又只能缓存get/load查询，结合自己的系统优化缓存模式还是很必要的，例如使用AOP进行Service切面缓存，这样就与DAO的ORM级别缓存分开了，相对更灵活。

上面是我对Hibernate和ehcache的一点了解，如果有描述不当或者错误的地方欢迎指正！