Details

    • Type: Bug
    • Status: Closed (View Workflow)
    • Priority: Major
    • Resolution: Done
    • Affects Version/s: 4.16.2
    • Fix Version/s: 4.17.0
    • Component/s: Testing, XTS
    • Labels:
      None

      Description

      I ran the crash recovery tests 3 times, they all hung at different times. It wasn't clear to me what the server was doing when it hung. For example, it wasn't being polled for a crash, or re-attempting recovery in a cycle. The problem can be recreated by running all the tests in one go:

      cd ./XTS/sar/crash-recovery-tests
      mvn test
      

      I tried running each JUnit test one at a time, and none of them hung. This can be done as follows:

      for i in $(ls src/test/java/com/arjuna/qa/junit/Test*); do TEST=$(echo $i | awk -F '.' '{ print $1 }' | sed 's/\//\./g' | cut -c 15-); mvn test -Dtest=$TEST; done
      cat target/surefire-reports/*.txt | grep "Tests run"
      

        Gliffy Diagrams

          Activity

          Hide
          tomjenkinson Tom Jenkinson added a comment -

          Hi Paul,

          I would like to put out a 4.16.3 start of next week, do you think that is feasible in so far as this issue is concerned?

          Tom

          Show
          tomjenkinson Tom Jenkinson added a comment - Hi Paul, I would like to put out a 4.16.3 start of next week, do you think that is feasible in so far as this issue is concerned? Tom
          Hide
          paul.robinson Paul Robinson added a comment - - edited

          We can push it back, as we have a work-around. I'm confident this is only a bug in the test, as we restart the server and ensure the object-store is empty between tests.

          Show
          paul.robinson Paul Robinson added a comment - - edited We can push it back, as we have a work-around. I'm confident this is only a bug in the test, as we restart the server and ensure the object-store is empty between tests.
          Hide
          tomjenkinson Tom Jenkinson added a comment -

          I am trying to decide whether to drop 4.x from the fix for and leave it as an "affects" as I don't especially want to plan another 4.x, wdyt?

          Show
          tomjenkinson Tom Jenkinson added a comment - I am trying to decide whether to drop 4.x from the fix for and leave it as an "affects" as I don't especially want to plan another 4.x, wdyt?
          Hide
          paul.robinson Paul Robinson added a comment -

          I suspect you may have fixed this when you fixed JBTM-1085. Can you check?

          Show
          paul.robinson Paul Robinson added a comment - I suspect you may have fixed this when you fixed JBTM-1085 . Can you check?
          Hide
          paul.robinson Paul Robinson added a comment -

          Brought forward to 5.0.0.M2 as it may now be fixed. If it's not fixed, we should move back to 5.0.0.Final.

          Show
          paul.robinson Paul Robinson added a comment - Brought forward to 5.0.0.M2 as it may now be fixed. If it's not fixed, we should move back to 5.0.0.Final.
          Hide
          zhfeng Amos Feng added a comment -

          I think this is fixed and I've backport to 4.16 branch.

          Show
          zhfeng Amos Feng added a comment - I think this is fixed and I've backport to 4.16 branch.
          Hide
          paul.robinson Paul Robinson added a comment -

          Great,

          Can you ensure the CI scripts use the following rather than the workaround.

          cd ./XTS/sar/crash-recovery-tests
          mvn test
          

          If CI then passes for trunk and branch, we can resolve this issue.

          Show
          paul.robinson Paul Robinson added a comment - Great, Can you ensure the CI scripts use the following rather than the workaround. cd ./XTS/sar/crash-recovery-tests mvn test If CI then passes for trunk and branch, we can resolve this issue.

            People

            • Assignee:
              zhfeng Amos Feng
              Reporter:
              paul.robinson Paul Robinson
            • Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:

                Development