头图

上篇文章提到查询时会用到缓存,其内置的两级缓存如下:

// 一级缓存,在executor中,与sqlsession绑定
// org.apache.ibatis.executor.BaseExecutor#localCache
// 指向org.apache.ibatis.cache.impl.PerpetualCache#cache
private Map<Object, Object> cache = new HashMap<>();

// 二级缓存,在MappedStatement中(对应mapper.xml中的一个crud方法),周期与SqlSessionFactory一致
org.apache.ibatis.mapping.MappedStatement#cache
// 最终也指向了org.apache.ibatis.cache.impl.PerpetualCache#cache
private Map<Object, Object> cache = new HashMap<>();
  • 一、二级缓存都是查询缓存,select写入,insert、update、delete则清除
  • 一、二级缓存均指向org.apache.ibatis.cache.impl.PerpetualCache#cache,本质是一个HashMap
  • 一、二级缓存Key的计算方式一致,均指向org.apache.ibatis.executor.BaseExecutor#createCacheKey,Key的本质:statement的id + offset + limit + sql + param参数
  • 一级缓存生命周期和SqlSession一致,默认开启;二级缓存声明周期和SqlSessionFactory一致,需手动开启
  • 相同namespace使用同一个二级缓存;二级缓存和事务关联,事务提交数据才会写入缓存,事务回滚则不会写入

接下来通过源码分别来看一下。

一级缓存

一级缓存的生命周期是sqlSession;在同一sqlSession中,用相同sql和查询条件多次查询DB情况,非首次查询会命中一级缓存。

一级缓存默认是开启的,如果想关闭需要增加配置

// == 如果不设置,默认是SESSION(后续的源码分析会涉及这里)
<setting name="localCacheScope" value="STATEMENT"/>

以查询方法作为入口

org.apache.ibatis.session.defaults.DefaultSqlSession#selectList(java.lang.String, java.lang.Object, org.apache.ibatis.session.RowBounds)
org.apache.ibatis.executor.BaseExecutor#query(org.apache.ibatis.mapping.MappedStatement, java.lang.Object, org.apache.ibatis.session.RowBounds, org.apache.ibatis.session.ResultHandler)
List<E> query(MappedStatement ms, Object parameter, RowBounds rowBounds, ResultHandler resultHandler) throws SQLException {
    BoundSql boundSql = ms.getBoundSql(parameter);
    // == 计算CacheKey
    CacheKey key = createCacheKey(ms, parameter, rowBounds, boundSql);
    // == 查询中使用缓存
    return query(ms, parameter, rowBounds, resultHandler, key, boundSql);
}

CacheKey计算

org.apache.ibatis.executor.BaseExecutor#createCacheKey
CacheKey createCacheKey(MappedStatement ms, Object parameterObject, RowBounds rowBounds, BoundSql boundSql) {
    CacheKey cacheKey = new CacheKey();
    // == 调用update方法修改cache
    cacheKey.update(ms.getId());
    cacheKey.update(rowBounds.getOffset());
    cacheKey.update(rowBounds.getLimit());
    cacheKey.update(boundSql.getSql());
    // value是参数
    cacheKey.update(value);
    return cacheKey;
}

从这里就可以猜测到,CacheKey和statement的id、offset、limit、sql、param参数有关

进入CacheKey验证这个猜测:

### CacheKey类 ###

// 默认37
private final int multiplier;
// 默认17
private int hashcode;
private long checksum;
private int count;

private List<Object> updateList;

public void update(Object object) {
    int baseHashCode = object == null ? 1 : ArrayUtil.hashCode(object);
    
    // -- 修改几个属性值
    count++;
    checksum += baseHashCode;
    baseHashCode *= count;
    hashcode = multiplier * hashcode + baseHashCode;
    // -- updateList新增对象
    updateList.add(object);
}

public boolean equals(Object object) {
    // -- 比较几个属性值
    if (hashcode != cacheKey.hashcode) {
      return false;
    }
    if (checksum != cacheKey.checksum) {
      return false;
    }
    if (count != cacheKey.count) {
      return false;
    }
    // -- 挨个比较updateList中的对象
    for (int i = 0; i < updateList.size(); i++) {
      Object thisObject = updateList.get(i);
      Object thatObject = cacheKey.updateList.get(i);
      if (!ArrayUtil.equals(thisObject, thatObject)) {
        return false;
      }
    }
    return true;
}

查询中使用缓存

org.apache.ibatis.executor.BaseExecutor#query(org.apache.ibatis.mapping.MappedStatement, java.lang.Object, org.apache.ibatis.session.RowBounds, org.apache.ibatis.session.ResultHandler, org.apache.ibatis.cache.CacheKey, org.apache.ibatis.mapping.BoundSql)
List<E> query(MappedStatement ms, Object parameter, RowBounds rowBounds, ResultHandler resultHandler, CacheKey key, BoundSql boundSql) throws SQLException {
    List<E> list;
    try {
      queryStack++;
      // == 1.先从localCache获取数据
      list = resultHandler == null ? (List<E>) localCache.getObject(key) : null;
      if (list != null) {
          handleLocallyCachedOutputParameters(ms, key, parameter, boundSql);
      } 
      // == 2.缓存中无数据,从数据库查询
      else {
          list = queryFromDatabase(ms, parameter, rowBounds, resultHandler, key, boundSql);
      }
    }

    // ## 如果scope设置成STATEMENT类型,会清理一级缓存
    if (configuration.getLocalCacheScope() == LocalCacheScope.STATEMENT) {
      // 清理缓存
      clearLocalCache();
    }
    return list;
}

继续观察代码2位置:

List<E> queryFromDatabase(MappedStatement ms, Object parameter, RowBounds rowBounds, ResultHandler resultHandler, CacheKey key, BoundSql boundSql) throws SQLException {
  List<E> list;
  // 缓存占位,表示正在执行
  localCache.putObject(key, EXECUTION_PLACEHOLDER);
  try {
    // == 查询DB逻辑
    list = doQuery(ms, parameter, rowBounds, resultHandler, boundSql);
  } finally {
    localCache.removeObject(key);
  }
  // == 执行结果放入一级缓存
  localCache.putObject(key, list);
  if (ms.getStatementType() == StatementType.CALLABLE) {
    localOutputParameterCache.putObject(key, parameter);
  }
  return list;
}

综上,查询过程会向localCache中存放查询结果。
只不过设置scope为STATEMENT时,每次都会清空缓存——这就是一级缓存失效的秘密

增删改清理缓存

insert和delete方法都会执行update:

public int insert(String statement) {
    return insert(statement, null);
}

public int delete(String statement) {
    return update(statement, null);
}

于是观察update即可:

int update(MappedStatement ms, Object parameter) throws SQLException {
    ErrorContext.instance().resource(ms.getResource()).activity("executing an update").object(ms.getId());
    if (closed) {
      throw new ExecutorException("Executor was closed.");
    }
    // == 清理一级缓存
    clearLocalCache();
    return doUpdate(ms, parameter);
}

二级缓存

二级缓存需要打开开关:

  • 第1步
<setting name="cacheEnabled" value="STATEMENT"/>
  • 第二步

同时在mapper.xml中增加标签

<cache/> 

默认的,二级缓存的key是namespace,如果要引用其它命名空间的Cache配置,可以使用如下标签:

<cache-ref namespace="xxx"/>

CachingExecutor

二级缓存的入口在executor创建位置:

public Executor newExecutor(Transaction transaction, ExecutorType executorType) {
    Executor executor;
    if (ExecutorType.BATCH == executorType) {
      executor = new BatchExecutor(this, transaction);
    } else if (ExecutorType.REUSE == executorType) {
      executor = new ReuseExecutor(this, transaction);
    } else {
      // 默认创建SimpleExecutor
      executor = new SimpleExecutor(this, transaction);
    }
    if (cacheEnabled) {
      // == 开启二级缓存情况,使用装饰器模式用CachingExecutor包了一层
      executor = new CachingExecutor(executor);
    }
    return executor;
}

观察构造器里做了什么

// 属性互相赋值
public CachingExecutor(Executor delegate) {
  this.delegate = delegate;
  delegate.setExecutorWrapper(this);
}

赋值后CachingExecutor和SimpleExecutor的关系如下
image.png

知道这一点后,我们来查看CachingExecutor的query方法:

public <E> List<E> query(MappedStatement ms, Object parameterObject, RowBounds rowBounds, ResultHandler resultHandler) throws SQLException {
    BoundSql boundSql = ms.getBoundSql(parameterObject);
    // == 调用delegate的createCacheKey方法(前面已经分析过)
    CacheKey key = createCacheKey(ms, parameterObject, rowBounds, boundSql);
    // == 二级缓存的查询
    return query(ms, parameterObject, rowBounds, resultHandler, key, boundSql);
}

观察query方法的实现

public <E> List<E> query(MappedStatement ms, Object parameterObject, RowBounds rowBounds, ResultHandler resultHandler, CacheKey key, BoundSql boundSql)
      throws SQLException {
    // ## A.通过MappedStatement获取cache
    Cache cache = ms.getCache();
    if (cache != null) {
      // 缓存刷新
      flushCacheIfRequired(ms);
      if (ms.isUseCache() && resultHandler == null) {
        ensureNoOutParams(ms, boundSql);
        // -- 1.通过tcm获取查询结果
        List<E> list = (List<E>) tcm.getObject(cache, key);
        if (list == null) {
          // -- 2.tcm中无结果,通过原executor查询(一级缓存+jdbc逻辑)
          list = delegate.query(ms, parameterObject, rowBounds, resultHandler, key, boundSql);
          // -- 3.查询结果最终放入tcm中
          tcm.putObject(cache, key, list); 
        }
        return list;
      }
    }
    return delegate.query(ms, parameterObject, rowBounds, resultHandler, key, boundSql);
}

// ## B.tcm指向这里
TransactionalCacheManager tcm = new TransactionalCacheManager();

整理逻辑很简单,但又有两个问题困扰到我

  1. 通过MappedStatement获取到的二级缓存cache(代码A位置),什么时候初始化的?
  2. 二级缓存和tcm(TransactionalCacheManager)之间有什么联系?

二级缓存初始化

沿着cache倒推,能追溯到Mapper解析。

完整调用链如下(当作复习了):

// 创建SqlSessionFactory
org.apache.ibatis.session.SqlSessionFactoryBuilder#build(java.io.Reader, java.lang.String, java.util.Properties)
org.apache.ibatis.builder.xml.XMLConfigBuilder#parse
// configuration解析
org.apache.ibatis.builder.xml.XMLConfigBuilder#parseConfiguration
// 解析mapper
org.apache.ibatis.builder.xml.XMLConfigBuilder#mapperElement
org.apache.ibatis.builder.xml.XMLMapperBuilder#parse
org.apache.ibatis.builder.xml.XMLMapperBuilder#configurationElement{
    // == 二级缓存的配置引用(执行namespace)
    cacheRefElement(context.evalNode("cache-ref"));
    // == 二级缓存的开启
    cacheElement(context.evalNode("cache"));
}

org.apache.ibatis.builder.xml.XMLMapperBuilder#cacheElement
org.apache.ibatis.builder.MapperBuilderAssistant#useNewCache{
    // == 二级缓存创建
    Cache cache = new CacheBuilder(currentNamespace)
        // -- Cache实现是PerpetualCache
        .implementation(valueOrDefault(typeClass, PerpetualCache.class))
        // -- 包装器用了LruCache
        .addDecorator(valueOrDefault(evictionClass, LruCache.class))
        .clearInterval(flushInterval)
        .size(size)
        .readWrite(readWrite)
        .blocking(blocking)
        .properties(props)
        .build();
}

看下二级缓存的整个装饰链(盗图)
image.png

SynchronizedCache -> LoggingCache -> SerializedCache -> LruCache -> PerpetualCache。

二级缓存和TransactionalCacheManager的关系

TransactionalCacheManager类:

// ## 维护一个map,key是Cache,value是TransactionalCache
Map<Cache, TransactionalCache> transactionalCaches = new HashMap<>();

public Object getObject(Cache cache, CacheKey key) {
    
           // ## 1.此方法会在transactionalCaches中建立k-v关系
    return getTransactionalCache(cache)
                ⬇⬇⬇⬇⬇⬇
                transactionalCaches.computeIfAbsent(cache, TransactionalCache::new);
            
            // == 2.从二级缓存中获取
            .getObject(key);
}

再观察TransactionalCache

// == 二级缓存
private final Cache delegate;
// == 二级缓存清理标记
private boolean clearOnCommit;

// ####   以下两个集合可以理解为用来存放临时数据   ####
// == 等事务提交时,需要加入二级缓存的对象
private final Map<Object, Object> entriesToAddOnCommit;
// == 二级缓存中不存在的对象key
private final Set<Object> entriesMissedInCache;

public void putObject(Object key, Object object) {
    // 对象记录到entriesToAddOnCommit中
    entriesToAddOnCommit.put(key, object);
}

public Object getObject(Object key) {
    // 从二级缓存获取
    Object object = delegate.getObject(key);
    if (object == null) {
      // 二级缓存中不存在,在entriesMissedInCache记录key
      entriesMissedInCache.add(key);
    }
}

这里能够看出,二级缓存和Transaction(事务)有很深的纠葛。
那么具体有什么纠葛?

  • 事务提交

观察TransactionManager的commit方法:

org.apache.ibatis.cache.TransactionalCacheManager#commit
org.apache.ibatis.cache.decorators.TransactionalCache#commit
public void commit() {
    // == 刷新对象
    flushPendingEntries();
}

private void flushPendingEntries() {
    for (Map.Entry<Object, Object> entry : entriesToAddOnCommit.entrySet()) {
      // == 对象从entriesToAddOnCommit刷新到二级缓存中
      delegate.putObject(entry.getKey(), entry.getValue());
    }
}

此处能证明,事务提交时对象从一个临时集合entriesToAddOnCommit刷新至二级缓存

  • 事务回滚

再观察回滚方法

org.apache.ibatis.cache.decorators.TransactionalCache#rollback
public void rollback() {
    unlockMissedEntries();
    // == 重置,将临时集合数据清理
    reset();
}

private void reset() {
  clearOnCommit = false;
  entriesToAddOnCommit.clear();
  entriesMissedInCache.clear();
}

附录

P6-P7知识合辑


青鱼
268 声望25 粉丝

山就在那里,每走一步就近一些