Uploaded image for project: 'HornetQ'
  1. HornetQ
  2. HORNETQ-716

Bridge queue is never stopped if server dies with kill -9

This issue belongs to an archived project. You can view it, but you can't modify it. Learn more

    XMLWordPrintable

Details

    • Bug
    • Status: Closed
    • Critical
    • Resolution: Done
    • 2.2.0.GA
    • 2.2.8.Final
    • Core
    • None

    Description

      To simulator a physical server power outage or crash, or a JVM crash, e.g. OutOfMemoryError, a JVM in a cluster is shutdown using "kill -9 ${PID}" during a test with significant load.

      JMS Messages are distributed from each cluster member to all other cluster members in a round robin fashion. It is observed that the bridge queues to the down server continue to collect messages. Indeed it is also observed that not all messages are getting delivered.

      The desired behavior is that as soon as a message is failed to be delivered due to connectivity issues, that the downed server be marked as unreachable and skipped during round robin processing until the connection issue is resolved. If the server is up, then connections should last a long time, thus almost all of the time, messaging will not be affected by this logic, only during an adverse condition.

      Attachments

        Activity

          People

            csuconic@redhat.com Clebert Suconic
            kisexton_jira Kirk Sexton (Inactive)
            Archiver:
            samahaja@redhat.com Sagar Mahajan

            Dates

              Created:
              Updated:
              Resolved:
              Archived:

              PagerDuty