Uploaded image for project: 'Hadoop Map/Reduce'
  1. Hadoop Map/Reduce
  2. MAPREDUCE-3897

capacity scheduler - maxActiveApplicationsPerUser calculation can be wrong

    XMLWordPrintableJSON

Details

    • Bug
    • Status: Resolved
    • Critical
    • Resolution: Fixed
    • 0.23.0
    • 0.23.2
    • mrv2
    • None

    Description

      The capacity scheduler calculates the maxActiveApplications and the maxActiveApplicationsPerUser based on the config yarn.scheduler.capacity.maximum-applications or default 10000.

      MaxActiveApplications = max ( ceil ( clusterMemory/minAllocation * maxAMResource% * absoluteMaxCapacity), 1)

      MaxActiveAppsPerUser = max( ceil (maxActiveApplicationsComputedAbove * (userLimit%/100) * userLimitFactor), 1)

      maxActiveApplications is already multiplied by the queue absolute MAXIMUM capacity, so if max capacity > capacity and if you have user limit factor 1 (which is the default) and only 1 user is running, that user will not be allowed to use over the queue capacity, so having it relative to MAX capacity doesn't make sense. That user could easily end up in a deadlock and all its space used by application masters.

      Attachments

        1. MAPREDUCE-3897-1.txt
          3 kB
          Eric Payne
        2. MAPREDUCE-3897-1.txt
          3 kB
          Eric Payne
        3. MAPREDUCE-3897-2.txt
          5 kB
          Eric Payne

        Activity

          People

            epayne Eric Payne
            tgraves Thomas Graves
            Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved: