Uploaded image for project: 'Apache Tez'
  1. Apache Tez
  2. TEZ-3222

Reduce messaging overhead for auto-reduce parallelism case

    XMLWordPrintableJSON

Details

    • Bug
    • Status: Closed
    • Major
    • Resolution: Fixed
    • None
    • 0.9.0
    • None
    • None

    Description

      A dag with 15k x 1000k vertex may auto-reduce to 15k x 1. And while the data size is appropriate for 1 task attempt, this results in an increase in task attempt message processing of 1000x.

      This jira aims to reduce the message processing in the auto-reduced task while keeping the amount of message processing in the AM the same or less.

      Attachments

        1. TEZ-3222.1.patch
          24 kB
          Jonathan Turner Eagles
        2. TEZ-3222.2.patch
          36 kB
          Jonathan Turner Eagles
        3. TEZ-3222.3.patch
          36 kB
          Jonathan Turner Eagles
        4. TEZ-3222.4.patch
          28 kB
          Jonathan Turner Eagles
        5. TEZ-3222.5.patch
          37 kB
          Jonathan Turner Eagles
        6. TEZ-3222.6.patch
          62 kB
          Jonathan Turner Eagles
        7. TEZ-3222.7.patch
          71 kB
          Jonathan Turner Eagles
        8. TEZ-3222.8.patch
          70 kB
          Jonathan Turner Eagles

        Issue Links

          blocks

          Sub-task - The sub-task of the issue TEZ-3361 Fetch Multiple Partitions from the Shuffle Handler

          • Major - Major loss of function.
          • Closed

          Activity

            People

              jeagles Jonathan Turner Eagles
              jeagles Jonathan Turner Eagles
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: