Uploaded image for project: 'Hadoop YARN'
  1. Hadoop YARN
  2. YARN-3933

FairScheduler: Multiple calls to completedContainer are not safe

    Details

    • Hadoop Flags:
      Reviewed

      Description

      In our cluster we are seeing available memory and cores being negative.
      Initial inspection:
      Scenario no. 1:
      In capacity scheduler the method allocateContainersToNode() checks if
      there are excess reservation of containers for an application, and they are no longer needed then it calls queue.completedContainer() which causes resources being negative. And they were never assigned in the first place.

      I am still looking through the code. Can somebody suggest how to simulate excess containers assignments ?

        Attachments

        1. YARN-3933.001.patch
          12 kB
          Shiwei Guo
        2. YARN-3933.002.patch
          12 kB
          Shiwei Guo
        3. YARN-3933.003.patch
          12 kB
          Shiwei Guo
        4. YARN-3933.004.patch
          14 kB
          Sunil Govindan
        5. YARN-3933.005.patch
          14 kB
          Sunil Govindan
        6. YARN-3933.006.patch
          6 kB
          Miklos Szegedi
        7. yarn-3933-branch-2.8.patch
          6 kB
          Karthik Kambatla

          Issue Links

            Activity

              People

              • Assignee:
                guoshiwei Shiwei Guo
                Reporter:
                lavkesh Lavkesh Lahngir
              • Votes:
                3 Vote for this issue
                Watchers:
                27 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved: