Update README.txt

README.txt

@@ -1,3 +1,7 @@
+
+本项目为本人近期阅读Hadoop源码时fork出来的，主要用于注释源码。
+
+
 For the latest information about Hadoop, please visit our website at:
 
    http://hadoop.apache.org/core/

hadoop-common-project/hadoop-common/pom.xml

@@ -616,7 +616,7 @@
                   <goal>javah</goal>
                 </goals>
                 <configuration>
-                  <javahPath>${env.JAVA_HOME}/bin/javah</javahPath>
+                  <javahPath>${java.home}/bin/javah</javahPath>
                   <javahClassNames>
                     <javahClassName>org.apache.hadoop.io.compress.zlib.ZlibCompressor</javahClassName>
                     <javahClassName>org.apache.hadoop.io.compress.zlib.ZlibDecompressor</javahClassName>

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapred/IndexRecord.java

@@ -22,9 +22,27 @@
 
 @InterfaceAudience.LimitedPrivate({"MapReduce"})
 @InterfaceStability.Unstable
+
+/**
+ * Spill文件索引 在相应Reducer的数据请求时快速定位到相应的partition。
+ * 一个Spill文件对应一个索引，索引存储专门分配的缓冲中（对应map输出的
+ * 环形Buffer）
+ */
 public class IndexRecord {
+
+  /**
+   * 起始偏移量（字节数）
+   */
   public long startOffset;
+
+  /**
+   * Partition数据原始长度（字节数）
+   */
   public long rawLength;
+
+  /**
+   * partition数据长度，如果压缩则算压缩后的长度（字节数）
+   */
   public long partLength;
 
   public IndexRecord() { }

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/Partitioner.java

@@ -37,6 +37,17 @@
  *
  * <p>Note: If you require your Partitioner class to obtain the Job's
  * configuration object, implement the {@link Configurable} interface.</p>
+ *
+ * Partitioner对map输出的key进行划分,决定key及其记录应该发往哪一个reducer.
+ *
+ * 即指定每一个key应该由哪个reducer来处理.
+ *
+ * 发往同一个reducer的所有key组成一个Partition.
+ *
+ * 如果作业只有一个reducer,则框架不会该作业创建Partitioner.
+ *
+ * 作业使用哪一个Partitioner由用户配置决定,Partitioner的逻辑中需要使用作业的配置信息,
+ * 可以通过实现Configurable接口访问配置信息.
  * 
  * @see Reducer
  */
@@ -50,6 +61,9 @@ public abstract class Partitioner<KEY, VALUE> {
    *   
    * <p>Typically a hash function on a all or a subset of the key.</p>
    *
+   * 每个Partition(Reducer)对应一个整数编号,该方法返回代码key所属的Partition
+   * 的编号. 传入的Partition总数即为作业的reducer总数.
+   *
    * @param key the key to be partioned.
    * @param value the entry value.
    * @param numPartitions the total number of partitions.

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/Reducer.java

@@ -30,7 +30,7 @@
 /** 
  * Reduces a set of intermediate values which share a key to a smaller set of
  * values.  
- * 
+ *
  * <p><code>Reducer</code> implementations 
  * can access the {@link Configuration} for the job via the 
  * {@link JobContext#getConfiguration()} method.</p>
@@ -114,7 +114,12 @@
  *   }
  * }
  * </pre></blockquote>
- * 
+ *
+ *
+ * Reducer逻辑实现的模板方法. 执行入口为run(). setup , reduce, cleanup 是3个模板方法,可以实现任意一个方法改变逻辑.
+ *
+ * 作业的配置信息可以通过Context的getConfigurable方法获取.
+ *
  * @see Mapper
  * @see Partitioner
  */

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/TaskType.java

@@ -23,6 +23,8 @@
 
 /**
  * Enum for map, reduce, job-setup, job-cleanup, task-cleanup task types.
+ *
+ * 任务类型
  */
 @InterfaceAudience.Public
 @InterfaceStability.Stable