Details

    • Target Version/s:
    • Hadoop Flags:
      Reviewed

      Description

      While examining RM stack traces on a busy cluster I noticed a pattern of AMs stuck waiting for the scheduler lock trying to call getTransferredContainers. The scheduler lock is highly contended, especially on a large cluster with many nodes heartbeating, and it would be nice if we could find a way to eliminate the need to grab this lock during this call. We've already done similar work during AM allocate calls to make sure they don't needlessly grab the scheduler lock, and it would be good to do so here as well, if possible.

        Attachments

        1. 0001-YARN-3136.patch
          5 kB
          Sunil Govindan
        2. 0002-YARN-3136.patch
          6 kB
          Sunil Govindan
        3. 0003-YARN-3136.patch
          7 kB
          Sunil Govindan
        4. 0004-YARN-3136.patch
          7 kB
          Sunil Govindan
        5. 0005-YARN-3136.patch
          12 kB
          Sunil Govindan
        6. 0006-YARN-3136.patch
          7 kB
          Sunil Govindan
        7. 0007-YARN-3136.patch
          8 kB
          Sunil Govindan
        8. 0008-YARN-3136.patch
          9 kB
          Sunil Govindan
        9. 0009-YARN-3136.patch
          7 kB
          Sunil Govindan
        10. 00010-YARN-3136.patch
          7 kB
          Sunil Govindan
        11. 00011-YARN-3136.patch
          7 kB
          Sunil Govindan
        12. 00012-YARN-3136.patch
          7 kB
          Sunil Govindan
        13. 00013-YARN-3136.patch
          7 kB
          Jian He
        14. YARN-3136.branch-2.7.patch
          8 kB
          Wangda Tan

          Activity

            People

            • Assignee:
              sunilg Sunil Govindan
              Reporter:
              jlowe Jason Lowe
            • Votes:
              0 Vote for this issue
              Watchers:
              15 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: