Uploaded image for project: 'Hadoop Map/Reduce'
  1. Hadoop Map/Reduce
  2. MAPREDUCE-5508

JobTracker memory leak caused by unreleased FileSystem objects in JobInProgress#cleanupJob

    XMLWordPrintableJSON

Details

    • Bug
    • Status: Resolved
    • Critical
    • Resolution: Fixed
    • 1-win, 1.2.1
    • 1-win, 1.3.0
    • jobtracker
    • None
    • Reviewed

    Description

      MAPREDUCE-5351 fixed a memory leak problem but introducing another filesystem object (see "tempDirFs") that is not properly released.

       JobInProgress#cleanupJob()

  void cleanupJob() {
...
          tempDirFs = jobTempDirPath.getFileSystem(conf);
          CleanupQueue.getInstance().addToQueue(
              new PathDeletionContext(jobTempDirPath, conf, userUGI, jobId));
...
 if (tempDirFs != fs) {
      try {
        fs.close();
      } catch (IOException ie) {
...
}

      Attachments

        1. MAPREDUCE-5508.patch
          2 kB
          Douma Fang
        2. MAPREDUCE-5508.1.patch
          3 kB
          Douma Fang
        3. MAPREDUCE-5508.2.patch
          4 kB
          Douma Fang
        4. MAPREDUCE-5508.3.patch
          4 kB
          Douma Fang
        5. CleanupQueue.java
          6 kB
          viswanathan
        6. JobInProgress.java
          126 kB
          viswanathan

        Issue Links

          is related to

          Bug - A problem which impairs or prevents the functions of the product. HDFS-5211 Race condition between DistributedFileSystem#close and FileSystem#close can cause return of a closed DistributedFileSystem instance from the FileSystem cache.

          • Major - Major loss of function.
          • Resolved

          Bug - A problem which impairs or prevents the functions of the product. MAPREDUCE-5351 JobTracker memory leak caused by CleanupQueue reopening FileSystem

          • Critical - Crashes, loss of data, severe memory leak.
          • Closed

          Improvement - An improvement or enhancement to an existing feature or task. HADOOP-9993 FileSystem#close JavaDoc should state that the implementation must be idempotent.

          • Major - Major loss of function.
          • Open

          Activity

            People

              xifang Douma Fang
              xifang Douma Fang
              Votes:
              0 Vote for this issue
              Watchers:
              8 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: