MAPREDUCE-6635. Unsafe long to int conversion in UncompressedSplitLineReader and IndexOutOfBoundsException. Contributed by Junping Du.

(cherry picked from commit c6f2d761d5)
(cherry picked from commit f1999fe275)

hadoop-mapreduce-project/CHANGES.txt

@@ -412,6 +412,9 @@ Release 2.8.0 - UNRELEASED
     MAPREDUCE-6616. Fail to create jobhistory file if there are some multibyte
     characters in the job name. (Kousuke Saruta via aajisaka)
 
+    MAPREDUCE-6635. Unsafe long to int conversion in UncompressedSplitLineReader
+    and IndexOutOfBoundsException. (Junping Du via vvasudev)
+
 Release 2.7.3 - UNRELEASED
 
   INCOMPATIBLE CHANGES
@@ -451,6 +454,9 @@ Release 2.7.3 - UNRELEASED
     MAPREDUCE-6191. Improve clearing stale state of Java serialization
                     testcase.  (Sam Liu via Eric Yang)
 
+    MAPREDUCE-6635. Unsafe long to int conversion in UncompressedSplitLineReader
+    and IndexOutOfBoundsException. (Junping Du via vvasudev)
+
 Release 2.7.2 - 2016-01-25
 
   INCOMPATIBLE CHANGES
@@ -741,6 +747,9 @@ Release 2.6.5 - UNRELEASED
     MAPREDUCE-6191. Improve clearing stale state of Java serialization
                     testcase.  (Sam Liu via Eric Yang)
 
+    MAPREDUCE-6635. Unsafe long to int conversion in UncompressedSplitLineReader
+    and IndexOutOfBoundsException. (Junping Du via vvasudev)
+
 Release 2.6.4 - 2016-02-11
 
   INCOMPATIBLE CHANGES

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

@@ -53,8 +53,11 @@ public class UncompressedSplitLineReader extends SplitLineReader {
       throws IOException {
     int maxBytesToRead = buffer.length;
     if (totalBytesRead < splitLength) {
-      maxBytesToRead = Math.min(maxBytesToRead,
+      long leftBytesForSplit = splitLength - totalBytesRead;
-                                (int)(splitLength - totalBytesRead));
+      // check if leftBytesForSplit exceed Integer.MAX_VALUE
+      if (leftBytesForSplit <= Integer.MAX_VALUE) {
+        maxBytesToRead = Math.min(maxBytesToRead, (int)leftBytesForSplit);
+      }
     }
     int bytesRead = in.read(buffer, 0, maxBytesToRead);
 

hadoop-mapreduce-project/hadoop-mapreduce-client/hadoop-mapreduce-client-core/src/test/java/org/apache/hadoop/mapred/TestLineRecordReader.java

@@ -110,6 +110,43 @@ public class TestLineRecordReader {
         numRecordsNoSplits, numRecordsFirstSplit + numRecordsRemainingSplits);
   }
 
+  private void testLargeSplitRecordForFile(Configuration conf,
+      long firstSplitLength, long testFileSize, Path testFilePath)
+      throws IOException {
+    conf.setInt(org.apache.hadoop.mapreduce.lib.input.
+        LineRecordReader.MAX_LINE_LENGTH, Integer.MAX_VALUE);
+    assertTrue("unexpected firstSplitLength:" + firstSplitLength,
+        testFileSize < firstSplitLength);
+    String delimiter = conf.get("textinputformat.record.delimiter");
+    byte[] recordDelimiterBytes = null;
+    if (null != delimiter) {
+      recordDelimiterBytes = delimiter.getBytes(Charsets.UTF_8);
+    }
+    // read the data without splitting to count the records
+    FileSplit split = new FileSplit(testFilePath, 0, testFileSize,
+        (String[])null);
+    LineRecordReader reader = new LineRecordReader(conf, split,
+        recordDelimiterBytes);
+    LongWritable key = new LongWritable();
+    Text value = new Text();
+    int numRecordsNoSplits = 0;
+    while (reader.next(key, value)) {
+      ++numRecordsNoSplits;
+    }
+    reader.close();
+
+    // count the records in the first split
+    split = new FileSplit(testFilePath, 0, firstSplitLength, (String[])null);
+    reader = new LineRecordReader(conf, split, recordDelimiterBytes);
+    int numRecordsFirstSplit = 0;
+    while (reader.next(key, value)) {
+      ++numRecordsFirstSplit;
+    }
+    reader.close();
+    assertEquals("Unexpected number of records in split",
+        numRecordsNoSplits, numRecordsFirstSplit);
+  }
+
   @Test
   public void testBzip2SplitEndsAtCR() throws IOException {
     // the test data contains a carriage-return at the end of the first
@@ -331,6 +368,22 @@ public class TestLineRecordReader {
     return file;
   }
 
+  @Test
+  public void testUncompressedInputWithLargeSplitSize() throws Exception {
+    Configuration conf = new Configuration();
+    // single char delimiter
+    String inputData = "abcde +fghij+ klmno+pqrst+uvwxyz";
+    Path inputFile = createInputFile(conf, inputData);
+    conf.set("textinputformat.record.delimiter", "+");
+    // split size over max value of integer
+    long longSplitSize = (long)Integer.MAX_VALUE + 1;
+    for (int bufferSize = 1; bufferSize <= inputData.length(); bufferSize++) {
+      conf.setInt("io.file.buffer.size", bufferSize);
+      testLargeSplitRecordForFile(conf, longSplitSize, inputData.length(),
+          inputFile);
+    }
+  }
+
   @Test
   public void testUncompressedInput() throws Exception {
     Configuration conf = new Configuration();