diff --git a/hadoop-mapreduce-project/CHANGES.txt b/hadoop-mapreduce-project/CHANGES.txt
index 9a1508beb83..18819ace67c 100644
--- a/hadoop-mapreduce-project/CHANGES.txt
+++ b/hadoop-mapreduce-project/CHANGES.txt
@@ -2072,6 +2072,9 @@ Release 0.22.0 - Unreleased
     MAPREDUCE-2991. queueinfo.jsp fails to show queue status for Capacity 
     scheduler if queue names contain special symbols. (Priyo Mustafi via shv)
 
+    MAPREDUCE-2779. JobSplitWriter.java can't handle large job.split file.
+    (Ming Ma via shv)
+
 Release 0.21.1 - Unreleased
 
   NEW FEATURES
diff --git a/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/split/JobSplitWriter.java b/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/split/JobSplitWriter.java
index 95f1fccfb7b..b6e44d71c45 100644
--- a/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/split/JobSplitWriter.java
+++ b/hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/main/java/org/apache/hadoop/mapreduce/split/JobSplitWriter.java
@@ -114,15 +114,15 @@ public class JobSplitWriter {
     if (array.length != 0) {
       SerializationFactory factory = new SerializationFactory(conf);
       int i = 0;
-      long offset = out.size();
+      long offset = out.getPos();
       for(T split: array) {
-        int prevCount = out.size();
+        long prevCount = out.getPos();
         Text.writeString(out, split.getClass().getName());
         Serializer<T> serializer = 
           factory.getSerializer((Class<T>) split.getClass());
         serializer.open(out);
         serializer.serialize(split);
-        int currCount = out.size();
+        long currCount = out.getPos();
         info[i++] = 
           new JobSplit.SplitMetaInfo( 
               split.getLocations(), offset,
@@ -139,12 +139,12 @@ public class JobSplitWriter {
     SplitMetaInfo[] info = new SplitMetaInfo[splits.length];
     if (splits.length != 0) {
       int i = 0;
-      long offset = out.size();
+      long offset = out.getPos();
       for(org.apache.hadoop.mapred.InputSplit split: splits) {
-        int prevLen = out.size();
+        long prevLen = out.getPos();
         Text.writeString(out, split.getClass().getName());
         split.write(out);
-        int currLen = out.size();
+        long currLen = out.getPos();
         info[i++] = new JobSplit.SplitMetaInfo( 
             split.getLocations(), offset,
             split.getLength());