MAPREDUCE-6165. [JDK8] TestCombineFileInputFormat failed on JDK8. Contributed by Akira AJISAKA.

(cherry picked from commit 551615fa13f65ae996bae9c1bacff189539b6557)
(cherry picked from commit 4e96175b334925466a9790e6ca20d0ec7d350791)

hadoop-mapreduce-project/CHANGES.txt

@@ -21,6 +21,9 @@ Release 2.7.5 - 2017-12-01
     MAPREDUCE-6957. shuffle hangs after a node manager connection timeout.
     (Jooseong Kim via jlowe)
 
+    MAPREDUCE-6165. [JDK8] TestCombineFileInputFormat failed on JDK8.
+    Contributed by Akira AJISAKA.
+
 Release 2.7.4 - 2017-08-04
 
   INCOMPATIBLE CHANGES

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/lib/input/CombineFileInputFormat.java

@@ -29,7 +29,6 @@ import java.util.HashMap;
 import java.util.Set;
 import java.util.Iterator;
 import java.util.Map;
-import java.util.Map.Entry;
 
 import org.apache.commons.logging.Log;
 import org.apache.commons.logging.LogFactory;
@@ -289,6 +288,26 @@ public abstract class CombineFileInputFormat<K, V>
                  maxSize, minSizeNode, minSizeRack, splits);
   }
 
+  /**
+   * Process all the nodes and create splits that are local to a node.
+   * Generate one split per node iteration, and walk over nodes multiple times
+   * to distribute the splits across nodes.
+   * <p>
+   * Note: The order of processing the nodes is undetermined because the
+   * implementation of nodeToBlocks is {@link java.util.HashMap} and its order
+   * of the entries is undetermined.
+   * @param nodeToBlocks Mapping from a node to the list of blocks that
+   *                     it contains.
+   * @param blockToNodes Mapping from a block to the nodes on which
+   *                     it has replicas.
+   * @param rackToBlocks Mapping from a rack name to the list of blocks it has.
+   * @param totLength Total length of the input files.
+   * @param maxSize Max size of each split.
+   *                If set to 0, disable smoothing load.
+   * @param minSizeNode Minimum split size per node.
+   * @param minSizeRack Minimum split size per rack.
+   * @param splits New splits created by this method are added to the list.
+   */
   @VisibleForTesting
   void createSplits(Map<String, Set<OneBlockInfo>> nodeToBlocks,
                      Map<OneBlockInfo, String[]> blockToNodes,
@@ -309,11 +328,6 @@ public abstract class CombineFileInputFormat<K, V>
     Set<String> completedNodes = new HashSet<String>();
     
     while(true) {
-      // it is allowed for maxSize to be 0. Disable smoothing load for such cases
-
-      // process all nodes and create splits that are local to a node. Generate
-      // one split per node iteration, and walk over nodes multiple times to
-      // distribute the splits across nodes. 
       for (Iterator<Map.Entry<String, Set<OneBlockInfo>>> iter = nodeToBlocks
           .entrySet().iterator(); iter.hasNext();) {
         Map.Entry<String, Set<OneBlockInfo>> one = iter.next();