Derby
  1. Derby
  2. DERBY-4037

On Z/OS storetests/st_reclaim_longcol.java fails with a large number of pages

    Details

    • Type: Bug Bug
    • Status: Closed
    • Priority: Major Major
    • Resolution: Cannot Reproduce
    • Affects Version/s: 10.3.3.0
    • Fix Version/s: None
    • Component/s: Store, Test
    • Environment:
    • Urgency:
      Normal
    • Bug behavior facts:
      Regression Test Failure

      Description

      On Z/OS 64 bit (haven't checked 32bit) st_reclaim_loncol fails with a large number of pages.

                      • Diff file derbyall/storeall/storetests/st_reclaim_longcol.diff
          • Start: st_reclaim_longcol jdk1.6.0 storeall:storetests 2009-01-23 14:20:49 ***
            3a4,6
            > Test failed, expected less than 150 pages - count is:
            > free pages : 258
            > allocated pages: 2047
            Test Failed.
          • End: st_reclaim_longcol jdk1.6.0 storeall:storetests 2009-01-23 14:25:07 ***

        Issue Links

          Activity

          Hide
          Knut Anders Hatlen added a comment -

          Triaged for 10.5.2. Duplicate of DERBY-1913.

          Show
          Knut Anders Hatlen added a comment - Triaged for 10.5.2. Duplicate of DERBY-1913 .
          Hide
          Kathey Marsden added a comment -

          I would like to reopen this issue because it fails consistently with an over 2000 page difference vs the small number of pages and an intermittent nature associated with DERBY-1913. I think maybe it is a different issue.

          Show
          Kathey Marsden added a comment - I would like to reopen this issue because it fails consistently with an over 2000 page difference vs the small number of pages and an intermittent nature associated with DERBY-1913 . I think maybe it is a different issue.
          Hide
          Knut Anders Hatlen added a comment -

          Feel free to reopen if you don't think they are duplicates.

          Show
          Knut Anders Hatlen added a comment - Feel free to reopen if you don't think they are duplicates.
          Hide
          Kathey Marsden added a comment -

          reopening the issue as I think it may not be a duplicate.

          Show
          Kathey Marsden added a comment - reopening the issue as I think it may not be a duplicate.
          Hide
          Kathey Marsden added a comment -

          Thanks Mike for the improved test. I tried it on z/os and still see this failure:

                          • Diff file storetests/storetests/st_reclaim_longcol.diff
              • Start: st_reclaim_longcol jdk1.6.0 storetests:storetests 2009-08-13 11:52:09 ***

          3a4,7
          > Test 2 failed, expected less than 150 pages - count is:
          > free pages : 2646
          > allocated pages: 55
          > Waited 100020ms. for background work.
          Test Failed.

              • End: st_reclaim_longcol jdk1.6.0 storetests:storetests 2009-08-13 11:55:50 ***

          This is with:
          java version "1.6.0"
          Java(TM) SE Runtime Environment (build pmz3160sr5-20090517_02(SR5))
          IBM J9 VM (build 2.4, J2RE 1.6.0 IBM J9 2.4 z/OS s390-31 jvmmz3160sr5-20090516_35588 (JIT enabled, AOT enabled)
          J9VM - 20090516_035588_bHdSMr
          JIT - r9_20090515_1756
          GC - 20090417_AA)
          JCL - 20090516_01

          Show
          Kathey Marsden added a comment - Thanks Mike for the improved test. I tried it on z/os and still see this failure: Diff file storetests/storetests/st_reclaim_longcol.diff Start: st_reclaim_longcol jdk1.6.0 storetests:storetests 2009-08-13 11:52:09 *** 3a4,7 > Test 2 failed, expected less than 150 pages - count is: > free pages : 2646 > allocated pages: 55 > Waited 100020ms. for background work. Test Failed. End: st_reclaim_longcol jdk1.6.0 storetests:storetests 2009-08-13 11:55:50 *** This is with: java version "1.6.0" Java(TM) SE Runtime Environment (build pmz3160sr5-20090517_02(SR5)) IBM J9 VM (build 2.4, J2RE 1.6.0 IBM J9 2.4 z/OS s390-31 jvmmz3160sr5-20090516_35588 (JIT enabled, AOT enabled) J9VM - 20090516_035588_bHdSMr JIT - r9_20090515_1756 GC - 20090417_AA) JCL - 20090516_01
          Hide
          Mike Matrigali added a comment -

          thanks for the new report. It looks like the test needs some more tweaking. The new output shows that the wait at end did allow for the system to free up a bunch of pages, vs. the previous
          report so I don't think there is a real problem here. But the test really wants the free'ing to come
          in the loop so that the subsequent insert can use the free'd pages of the previously committed
          deleted. For some reason this machine must not allow as concurrent background thread work
          as other platforms. I'll think about how best to change the test.

          Show
          Mike Matrigali added a comment - thanks for the new report. It looks like the test needs some more tweaking. The new output shows that the wait at end did allow for the system to free up a bunch of pages, vs. the previous report so I don't think there is a real problem here. But the test really wants the free'ing to come in the loop so that the subsequent insert can use the free'd pages of the previously committed deleted. For some reason this machine must not allow as concurrent background thread work as other platforms. I'll think about how best to change the test.
          Hide
          Mike Matrigali added a comment -

          triaged for 10.8.

          Show
          Mike Matrigali added a comment - triaged for 10.8.
          Hide
          Kathey Marsden added a comment -

          In the platform testing for 10.8 http://wiki.apache.org/db-derby/TenEightOnePlatformTesting
          This did not occur in any of the four JDK's tested. These have all had updates since the
          last round of z/OS testing, so I am not sure yet if it was a JVM fix or a Derby change that
          caused it to start passing. Here for reference are the four versions that passed with 10.8.

          java version "1.5.0"
          Java(TM) 2 Runtime Environment, Standard Edition (build pmz31devifx-20110214 (SR
          12 FP4 ))
          IBM J9 VM (build 2.3, J2RE 1.5.0 IBM J9 2.3 z/OS s390-31 j9vmmz3123ifx-20110209
          (JIT enabled)
          J9VM - 20110202_74536_bHdSMr
          JIT - 20100623_16197ifx3_r8
          GC - 20100211_AA)
          JCL - 20110214
          $

          $ J5.0_64/bin/java -version
          java version "1.5.0"
          Java(TM) 2 Runtime Environment, Standard Edition (build pmz64devifx-20110214 (SR
          12 FP4 ))
          IBM J9 VM (build 2.3, J2RE 1.5.0 IBM J9 2.3 z/OS s390x-64 j9vmmz6423ifx-20110209
          (JIT enabled)
          J9VM - 20110202_74536_BHdSMr
          JIT - 20100623_16197ifx3_r8
          GC - 20100211_AA)
          JCL - 20110214

          java version "1.6.0"
          Java(TM) SE Runtime Environment (build pmz3160sr9fp1-20110303_01(SR9 FP1))
          IBM J9 VM (build 2.4, JRE 1.6.0 IBM J9 2.4 z/OS s390-31 jvmmz3160sr9-20110203_74
          623 (JIT enabled, AOT enabled)
          J9VM - 20110203_074623
          JIT - r9_20101028_17488ifx3
          GC - 20101027_AA)
          JCL - 20110203_01

          java version "1.6.0"
          Java(TM) SE Runtime Environment (build pmz6460sr9fp1-20110303_01(SR9 FP1))
          IBM J9 VM (build 2.4, JRE 1.6.0 IBM J9 2.4 z/OS s390x-64 jvmmz6460sr9-20110203_7
          4623 (JIT enabled, AOT enabled)
          J9VM - 20110203_074623
          JIT - r9_20101028_17488ifx3
          GC - 20101027_AA)
          JCL - 20110203_01

          Show
          Kathey Marsden added a comment - In the platform testing for 10.8 http://wiki.apache.org/db-derby/TenEightOnePlatformTesting This did not occur in any of the four JDK's tested. These have all had updates since the last round of z/OS testing, so I am not sure yet if it was a JVM fix or a Derby change that caused it to start passing. Here for reference are the four versions that passed with 10.8. java version "1.5.0" Java(TM) 2 Runtime Environment, Standard Edition (build pmz31devifx-20110214 (SR 12 FP4 )) IBM J9 VM (build 2.3, J2RE 1.5.0 IBM J9 2.3 z/OS s390-31 j9vmmz3123ifx-20110209 (JIT enabled) J9VM - 20110202_74536_bHdSMr JIT - 20100623_16197ifx3_r8 GC - 20100211_AA) JCL - 20110214 $ $ J5.0_64/bin/java -version java version "1.5.0" Java(TM) 2 Runtime Environment, Standard Edition (build pmz64devifx-20110214 (SR 12 FP4 )) IBM J9 VM (build 2.3, J2RE 1.5.0 IBM J9 2.3 z/OS s390x-64 j9vmmz6423ifx-20110209 (JIT enabled) J9VM - 20110202_74536_BHdSMr JIT - 20100623_16197ifx3_r8 GC - 20100211_AA) JCL - 20110214 java version "1.6.0" Java(TM) SE Runtime Environment (build pmz3160sr9fp1-20110303_01(SR9 FP1)) IBM J9 VM (build 2.4, JRE 1.6.0 IBM J9 2.4 z/OS s390-31 jvmmz3160sr9-20110203_74 623 (JIT enabled, AOT enabled) J9VM - 20110203_074623 JIT - r9_20101028_17488ifx3 GC - 20101027_AA) JCL - 20110203_01 java version "1.6.0" Java(TM) SE Runtime Environment (build pmz6460sr9fp1-20110303_01(SR9 FP1)) IBM J9 VM (build 2.4, JRE 1.6.0 IBM J9 2.4 z/OS s390x-64 jvmmz6460sr9-20110203_7 4623 (JIT enabled, AOT enabled) J9VM - 20110203_074623 JIT - r9_20101028_17488ifx3 GC - 20101027_AA) JCL - 20110203_01
          Hide
          Mike Matrigali added a comment -

          With the latest Z/OS jvm upgrades this problem no longer reproduces, so likely
          a JVM/scheduling issue. closing as can't reproduce. Feel free to reopen or log a new one if it reproduces again in the future with new JVM's.

          Show
          Mike Matrigali added a comment - With the latest Z/OS jvm upgrades this problem no longer reproduces, so likely a JVM/scheduling issue. closing as can't reproduce. Feel free to reopen or log a new one if it reproduces again in the future with new JVM's.
          Hide
          Knut Anders Hatlen added a comment -

          [bulk update] Close all resolved issues that haven't been updated for more than one year.

          Show
          Knut Anders Hatlen added a comment - [bulk update] Close all resolved issues that haven't been updated for more than one year.

            People

            • Assignee:
              Mike Matrigali
              Reporter:
              Kathey Marsden
            • Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:

                Development