Uploaded image for project: 'Hadoop YARN'
  1. Hadoop YARN
  2. YARN-45 [Preemption] Scheduler feedback to AM to release containers
  3. YARN-352

Inconsistent picture of how a container was killed when querying RM and NM in case of preemption

    XMLWordPrintableJSON

    Details

    • Type: Sub-task
    • Status: Open
    • Priority: Major
    • Resolution: Unresolved
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None

      Description

      When the RM preempts a container, it records the exit status as -100. However, at the NM, it registers the preempted container's exit status as simply killed by an external via SIGTERM or SIGKILL.

      When the AM queries the RM and NM for the same container's status, it will get 2 different values.

      When killing a container, the exit reason should likely be more defined via an exit status code for the AM to act on in addition to providing of the diagnostic messages that can contain more detailed information ( though probably not programmatically interpret-able by the AM ).

        Attachments

          Issue Links

            Activity

              People

              • Assignee:
                Unassigned
                Reporter:
                hitesh Hitesh Shah
              • Votes:
                0 Vote for this issue
                Watchers:
                9 Start watching this issue

                Dates

                • Created:
                  Updated: