[HADOOP-15088] BufferedInputStream.skip function can return 0 when the file is corrupted, causing an infinite loop

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 2.5.0
• Component/s: streaming
• Assignee: Unassigned
• Reporter: John Doe
• Created: 04/Dec/17 04:03
• Updated: 04/Dec/17 14:29

Description

When a file is corrupted, the skip function can return 0, causing an infinite loop.
Here is the code:

  private boolean slowReadUntilMatch(Pattern markPattern, boolean includePat,
                                     DataOutputBuffer outBufOrNull) throws IOException {
      ...
      for (long skiplen = endPos; skiplen > 0; ) {
        skiplen -= bin_.skip(skiplen); 
      }
      ...
  }

Similar bugs are Hadoop-8614, Yarn-2905, Yarn-163, MAPREDUCE-6990