ActivationGroup should have a more orderly shutdown

XMLWordPrintable

    • Type: Bug
    • Resolution: Duplicate
    • Priority: P4
    • None
    • Affects Version/s: 1.3.0
    • Component/s: core-libs
    • None
    • sparc
    • solaris_2.6

      Under certain conditions, the ActivationGroup will forcibly unexport itself in
      the midst of servicing a remote call which may cause the VM to exit before
      the call returns.

      One particular example is that if the first object activated in an
      ActivationGroup fails, the ActivationGroup will unexport itself, potentially
      causing the VM to exit before the "newInstance" call returns. In
      the finally clause of ActivationGroupImpl.newInstance, the checkInactiveGroup
      call will result in the ActivationGroup being forcibly unexported since
      there will be no objects in the activation group if the first activation
      in the group fails.

      The ActivationGroup should not forcibly unexport itself while executing a
      remote call. Also, the VM should not exit while remote calls are pending or
      are in progress (see 4308492).

            Assignee:
            Ann Wollrath (Inactive)
            Reporter:
            Ann Wollrath (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            0 Start watching this issue

              Created:
              Updated:
              Resolved:
              Imported:
              Indexed: