glassfish
  1. glassfish
  2. GLASSFISH-10392

Warning messages when shutting down instance with LOCAL broker and active connections

    Details

    • Type: Bug Bug
    • Status: Open
    • Priority: Major Major
    • Resolution: Unresolved
    • Affects Version/s: v2.1.1
    • Fix Version/s: None
    • Component/s: jca
    • Labels:
      None
    • Environment:

      Operating System: All
      Platform: PC

    • Issuezilla Id:
      10,392
    • Status Whiteboard:
      Hide

      v3_exclude 3.1-exclude

      Show
      v3_exclude 3.1-exclude
    • Tags:

      Description

      Configure Glassfish to use a LOCAL broker and run an application which creates
      and closes some outbound MQ connections (e.g. using a Session Bean). Then shut
      down the instance. The server log shows numerous warning messages.

      What follows is an example, see the attached server.log for a full example.

      MQJMSRA_CL1101: onEvent:Connection Event for mc=7 :xacId=:event:E201:[E201]:
      Connection closed due to admin requested shutdown: localhost:7676(2669),
      com.sun.messaging.jms.notification.ConnectionClosedEvent[source=BrokerAddress=localhost:7676(2669),
      ConnectionID=910048231024248064, ReconnectEnabled: true, IsConnectedToHABroker:
      false]

      [I500]: Caught JVM Exception: java.io.EOFException: Trying to read 72 bytes.
      Already read 0 bytes.

      [I107]: Connection recover state: RECOVER_INACTIVE, broker: localhost:7676(2669)

      [I107]: Connection recover state: RECOVER_INACTIVE, broker: localhost:7676(2669)
      [I107]: Connection recover state: RECOVER_INACTIVE, broker: localhost:7676(2669)
      [I500]: Caught JVM Exception: java.lang.NullPointerException
      [I500]: Caught JVM Exception: com.sun.messaging.jms.JMSException: [C4062]:
      Cannot perform operation, connection is closed.
      [I500]: Caught JVM Exception: java.net.SocketException: Socket is closed

      It appears that the problem is that the RA is being shut down before the
      connection pool is shut down.

      This problem can be seen with Glassfish 2.1.1 build 29.
      This problem does not occur with Glassfish V3 build 64.

        Activity

        Hide
        Nigel Deakin added a comment -

        Created an attachment (id=3550)
        server log to demonstrate issue

        Show
        Nigel Deakin added a comment - Created an attachment (id=3550) server log to demonstrate issue
        Hide
        Satish Kumar added a comment -

        Adding v3_exclude to the status whiteboard since this works on V3...

        Show
        Satish Kumar added a comment - Adding v3_exclude to the status whiteboard since this works on V3...
        Hide
        Satish Kumar added a comment -

        In v2.x, ra.stop is called before terminating all connections. Hence the
        exceptions below.

        This issue has been fixed in V3 but is still a problem in the v2.x code base.
        Requestion Jagadish to investigate this issue...

        Show
        Satish Kumar added a comment - In v2.x, ra.stop is called before terminating all connections. Hence the exceptions below. This issue has been fixed in V3 but is still a problem in the v2.x code base. Requestion Jagadish to investigate this issue...
        Hide
        Jagadish added a comment -

        This issue is not applicable for v3.x versions.
        v3.x versions has a proper shutdown sequence, all pools, resources are destroyed
        and then resource-adapter is stopped during server shutdown or rar undeployment.

        Show
        Jagadish added a comment - This issue is not applicable for v3.x versions. v3.x versions has a proper shutdown sequence, all pools, resources are destroyed and then resource-adapter is stopped during server shutdown or rar undeployment.
        Hide
        Jagadish added a comment -

        Fixing status whiteboard keys

        Show
        Jagadish added a comment - Fixing status whiteboard keys
        Hide
        jthoennes added a comment -

        Hi Jagadish,

        I am wondering why you assigned the milestone v2.1.2 to this issue.
        I was thinking the v2.1 development has been stopped to concentrate on v3.

        Did Sun Oracle change its mind?

        Thanks, Jörg

        Show
        jthoennes added a comment - Hi Jagadish, I am wondering why you assigned the milestone v2.1.2 to this issue. I was thinking the v2.1 development has been stopped to concentrate on v3. Did Sun Oracle change its mind? Thanks, Jörg
        Hide
        Jagadish added a comment -

        Hi Jorg,
        v2.x code base and v3.x code base are significantly different.
        The shutdown sequence of connection-pools and resource-adapter are proper in v3
        and hence this issue is not applicable for v3.x versions.
        This issue will be applicable for v2.x versions only but the default target
        milestone for any issue created in issue tracker is 3.0.1. Usually, the target
        milestone will be current development version (or next release) of the product.
        For 2.1.1 based code base, next logical release number available in issue
        tracker is 2.1.2 (though there are no current plans for 2.1.2, AFAIK) and hence
        I marked target milestone as 2.1.2
        If there is any other release number (> 2.1.1, based on v2.x code base) that
        will be planned as next release, I shall update the target milestone accordingly.

        Thanks,
        -Jagadish

        Show
        Jagadish added a comment - Hi Jorg, v2.x code base and v3.x code base are significantly different. The shutdown sequence of connection-pools and resource-adapter are proper in v3 and hence this issue is not applicable for v3.x versions. This issue will be applicable for v2.x versions only but the default target milestone for any issue created in issue tracker is 3.0.1. Usually, the target milestone will be current development version (or next release) of the product. For 2.1.1 based code base, next logical release number available in issue tracker is 2.1.2 (though there are no current plans for 2.1.2, AFAIK) and hence I marked target milestone as 2.1.2 If there is any other release number (> 2.1.1, based on v2.x code base) that will be planned as next release, I shall update the target milestone accordingly. Thanks, -Jagadish
        Hide
        Jagadish added a comment -

        3.1-exclude keyword

        Show
        Jagadish added a comment - 3.1-exclude keyword

          People

          • Assignee:
            Jagadish
            Reporter:
            Nigel Deakin
          • Votes:
            1 Vote for this issue
            Watchers:
            0 Start watching this issue

            Dates

            • Created:
              Updated: