glassfish
  1. glassfish
  2. GLASSFISH-14094

[BLOCKING] TX checkpoint failed on a cluster env with multiple machine config

    Details

    • Type: Bug Bug
    • Status: Resolved
    • Priority: Critical Critical
    • Resolution: Fixed
    • Affects Version/s: 3.1
    • Fix Version/s: 3.1_ms06
    • Component/s: failover
    • Labels:
      None
    • Environment:

      Operating System: All
      Platform: Linux

      Description

      Promote build 24.

      The simple transaction checkpoint test SFSBDriver passed on the configuration of
      a cluster with 4 instances on single machine. It failed on the configuration of
      a cluster with 8 instances on multiple machines.

      From the output, the JSESSIONID was reset after the failover:
      1. Before failover:
      ------------- WebConversation ------------
      Cookies =
      JSESSIONID : 6714dfc17936571a62bbf9026029
      JROUTE : P3iG
      JSESSIONIDVERSION : /SFSBDriver:4
      JREPLICA : instance107
      ------------- Request ------------

      2. After failover
      ------------- WebConversation ------------
      Cookies =
      JREPLICA : instance107
      JSESSIONID : 67197f813d0d0482a03bb6134b71
      JSESSIONIDVERSION : /SFSBDriver:4
      JROUTE : 9ADs
      ------------- Request ------------

      1. SFSBDriver.war
        12 kB
        mzh777
      2. testTXCPDriver.zip
        142 kB
        mzh777

        Issue Links

          Activity

          Hide
          mzh777 added a comment -

          Created an attachment (id=5178)
          The testing app.

          Show
          mzh777 added a comment - Created an attachment (id=5178) The testing app.
          Hide
          mzh777 added a comment -

          Tried to set the log level to
          asadmin set-log-levels --target=... org.shoal.ha.cache.command.load_request=FINE
          asadmin set-log-levels --target=... org.shoal.ha.cache.command.load_response=FINE
          asadmin set-log-levels --target=... org.shoal.ha.*=FINE

          There are still no enough info to isolate the issue. There seem to be some log
          issue here also. See the attached logs.

          Show
          mzh777 added a comment - Tried to set the log level to asadmin set-log-levels --target=... org.shoal.ha.cache.command.load_request=FINE asadmin set-log-levels --target=... org.shoal.ha.cache.command.load_response=FINE asadmin set-log-levels --target=... org.shoal.ha.*=FINE There are still no enough info to isolate the issue. There seem to be some log issue here also. See the attached logs.
          Hide
          mzh777 added a comment -

          Created an attachment (id=5179)
          Test logs.

          Show
          mzh777 added a comment - Created an attachment (id=5179) Test logs.
          Hide
          mzh777 added a comment -

          Mark it as blocking since most EJB checkpoint tests (70%+) are failing in this
          instances on multiple machine configuration.

          Show
          mzh777 added a comment - Mark it as blocking since most EJB checkpoint tests (70%+) are failing in this instances on multiple machine configuration.
          Hide
          Mahesh Kannan added a comment -

          We now have a patch for 13741. We will run this test using the patch tomorrow and
          update the issue

          Show
          Mahesh Kannan added a comment - We now have a patch for 13741. We will run this test using the patch tomorrow and update the issue
          Hide
          Mahesh Kannan added a comment -

          Ran with the new shoal jars and 14 out of 61 tests failed. So more than 75%
          passed.

          Show
          Mahesh Kannan added a comment - Ran with the new shoal jars and 14 out of 61 tests failed. So more than 75% passed.

            People

            • Assignee:
              Mahesh Kannan
              Reporter:
              mzh777
            • Votes:
              0 Vote for this issue
              Watchers:
              0 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved: