Derby
  1. Derby
  2. DERBY-5209

10.8 upgrade and compatibility testing

    Details

    • Type: Task Task
    • Status: Closed
    • Priority: Major Major
    • Resolution: Fixed
    • Affects Version/s: 10.8.2.2
    • Fix Version/s: 10.8.2.2
    • Component/s: Test
    • Labels:
      None

      Description

      I am doing some upgrade, client/server and test compatibility testing with 10.8 and am using this issue for tracking tests and results.
      The first test was to create the system/wombat database with 10.5 and run Suites.All with 10.8 product jars and 10.5 derbyTesting.jar Then I verified I could reconnect with 10.5 and perform a few operations.

      derbyTesting.jar. Environment for reference is:
      export JAVA_HOME="/ibmsvn/ntsoftware/ibm16"
      export PATH="$JAVA_HOME/bin:$PATH"
      LIB="c:/cygwin/home/kmarsden/projects/10.8.1testing/db-derby-10.8.1.2-lib/lib"
      TOOLS="C:/cygwin/svn4/trunka/tools/java"
      LIB_105="C:/cygwin/svnreleases/jars/10.5.3.0"
      export CLASSPATH="$LIB/derbyrun.jar;$LIB_105/derbyTesting.jar;$TOOLS/junit.jar;
      TOOLS/jakarta-oro-2.0.8.jar"

      Output attached as test10_8SoftUpgradeFrom10_5.out. There were Failures: 188, Errors: 109 which haven't been analyzed yet, but I skimmed them and didn't see anything that looked like data corruption.

      Will also run
      10.7 derbyclient.jar, derbyTesting.jar 10.8 other product jars

      10.7 derbynet.jar derbyTesting.jar 10.8 other product jars.

      1. test10_8SoftUpgradeFrom10_5.out
        1.44 MB
        Kathey Marsden
      2. 10_5SoftupgradeTestOutput.txt
        1.46 MB
        Kathey Marsden
      3. 10_5SoftupgradeNotes.txt
        11 kB
        Kathey Marsden
      4. 10_5SoftUpgradeAnd10_5clientTestOutput.txt
        1.36 MB
        Kathey Marsden

        Issue Links

          Activity

          Hide
          Kathey Marsden added a comment -

          Attaching output from 10.8 testing with 10.5 soft upgraded database. It has not been analyzed yet. Also notable was that with the IBM jvm 144 user generated javacore files were generated.

          Show
          Kathey Marsden added a comment - Attaching output from 10.8 testing with 10.5 soft upgraded database. It has not been analyzed yet. Also notable was that with the IBM jvm 144 user generated javacore files were generated.
          Hide
          Myrna van Lunteren added a comment -

          I haven't looked but in re DERBY-5028 I ran into some javacores being generated.
          Knut said:
          "Perhaps you see the javacore files because of DERBY-4856? Since the interrupts cause session severity errors, I think that code will kick in and generate thread dumps."
          And Lilly said:
          "There are two way to avoid the problem: deleting them to the cleanup or set the derby.stream.error.extendedDiagSeverityLevel higher so javacore file does not get created."

          Perhaps it's related.

          Show
          Myrna van Lunteren added a comment - I haven't looked but in re DERBY-5028 I ran into some javacores being generated. Knut said: "Perhaps you see the javacore files because of DERBY-4856 ? Since the interrupts cause session severity errors, I think that code will kick in and generate thread dumps." And Lilly said: "There are two way to avoid the problem: deleting them to the cleanup or set the derby.stream.error.extendedDiagSeverityLevel higher so javacore file does not get created." Perhaps it's related.
          Hide
          Lily Wei added a comment -

          Yes, if the test has encounter error, the fix for DERBY-4856 will generate javacore file. Since we don't have extenededDiagSeverityLevel test yet, it is worth investigate to see whether we can make this case as a extenedDiagSeverityLevel test. Thanks!

          Show
          Lily Wei added a comment - Yes, if the test has encounter error, the fix for DERBY-4856 will generate javacore file. Since we don't have extenededDiagSeverityLevel test yet, it is worth investigate to see whether we can make this case as a extenedDiagSeverityLevel test. Thanks!
          Hide
          Kathey Marsden added a comment -

          Attaching My notes and test output for soft upgrade testing from 10.5 with 10.8.2.1
          10_5SoftupgradeNotes.txt and 10_5SoftupgradeTestOutput.txt

          Created database with 10.5.3.0. Ran tests with 10.8.2.1 RC1 product jars and 10.5 derbyTesting.jar against a soft upgraded database created with 10.5.3.0 in order to identify potential soft upgrade impact and behavior changes that should be documented.

          I ran with IBM JDK 1.6
          $ java -version
          java version "1.6.0"
          Java(TM) SE Runtime Environment (build pwi3260sr9fp1-20110208_03(SR9 FP1))
          IBM J9 VM (build 2.4, JRE 1.6.0 IBM J9 2.4 Windows XP x86-32 jvmwi3260sr9-20110203_74623 (JIT enabled, AOT enabled)
          J9VM - 20110203_074623
          JIT - r9_20101028_17488ifx3
          GC - 20101027_AA)
          JCL - 20110203_01

          There were 192 Failures 109 Errors in the following categories:

          The most serious issue is DERBY-5427 which I filed, but I also need to file issues for the issues in the notes marked with asterisks

          *3) NullPointerException in AuthentcationTest.testSystemShutdown(). Did not reproduce on rerun of the individual test.

          testSystemShutdown Exception in thread "DRDAConnThread_898" java.lang.NullPointerException
          at org.apache.derby.iapi.services.monitor.Monitor.getStream(Unknown Source)
          at org.apache.derby.iapi.services.monitor.Monitor.logMessage(Unknown Source)
          at org.apache.derby.impl.drda.NetworkServerControlImpl.consoleMessage(Unknown Source)
          at org.apache.derby.impl.drda.NetworkServerControlImpl.consoleExceptionPrintTrace(Unknown Source)
          at org.apache.derby.impl.drda.DRDAConnThread.sendUnexpectedException(Unknown Source)
          at org.apache.derby.impl.drda.DRDAConnThread.handleException(Unknown Source)
          at org.apache.derby.impl.drda.DRDAConnThread.run(Unknown Source)
          used 1500 ms .

          **10) testFunctionNullHandling(org.apache.derbyTesting.functionTests.tests.lang.RoutineTest)junit.framework.AssertionFailedError: Column value mismatch @ column '1', row 1:
          Expected: >11:00:00<
          Found: >null<
          at org.apache.derbyTesting.junit.JDBC.assertRowInResultSet(JDBC.java:109...
          at org.apache.derbyTesting.functionTests.tests.lang.RoutineTest.testFunctionNullHandling(RoutineTest.java:575)
          Needs investigation. I don't see a specific change to the assertion since 10.5.

          *13) testUpdateXXXWithAllDatatypes(org.apache.derbyTesting.functionTests.tests.lang.UpdatableResultSetTest)junit.framework.AssertionFailedError: The expected SQL state must be five characters long caused by
          at org.apache.derbyTesting.functionTests.tests.lang.UpdatableResultSetTest.runTestUpdateXXXWithAllDatatypes(UpdatableResultSetTest.java:2726)
          Caused by: java.sql.SQLException: Stream or LOB value cannot be retrieved more than once
          The new error is expected but I am surprised that the SQLState would not be five characters. Needs Investigation.

          *22) Needs investigation. Test has not changed in years.
          testGreekCharacters(org.apache.derbyTesting.functionTests.tests.jdbcapi.DriverMgrAuthenticationTest)junit.framework.AssertionFailedError: Connection should've been refused/failed
          at org.apache.derbyTesting.functionTests.tests.jdbcapi.DriverMgrAuthenticationTest.assertConnectionFail(DriverMgrAuthenticationTest.java:116)

          Show
          Kathey Marsden added a comment - Attaching My notes and test output for soft upgrade testing from 10.5 with 10.8.2.1 10_5SoftupgradeNotes.txt and 10_5SoftupgradeTestOutput.txt Created database with 10.5.3.0. Ran tests with 10.8.2.1 RC1 product jars and 10.5 derbyTesting.jar against a soft upgraded database created with 10.5.3.0 in order to identify potential soft upgrade impact and behavior changes that should be documented. I ran with IBM JDK 1.6 $ java -version java version "1.6.0" Java(TM) SE Runtime Environment (build pwi3260sr9fp1-20110208_03(SR9 FP1)) IBM J9 VM (build 2.4, JRE 1.6.0 IBM J9 2.4 Windows XP x86-32 jvmwi3260sr9-20110203_74623 (JIT enabled, AOT enabled) J9VM - 20110203_074623 JIT - r9_20101028_17488ifx3 GC - 20101027_AA) JCL - 20110203_01 There were 192 Failures 109 Errors in the following categories: The most serious issue is DERBY-5427 which I filed, but I also need to file issues for the issues in the notes marked with asterisks *3) NullPointerException in AuthentcationTest.testSystemShutdown(). Did not reproduce on rerun of the individual test. testSystemShutdown Exception in thread "DRDAConnThread_898" java.lang.NullPointerException at org.apache.derby.iapi.services.monitor.Monitor.getStream(Unknown Source) at org.apache.derby.iapi.services.monitor.Monitor.logMessage(Unknown Source) at org.apache.derby.impl.drda.NetworkServerControlImpl.consoleMessage(Unknown Source) at org.apache.derby.impl.drda.NetworkServerControlImpl.consoleExceptionPrintTrace(Unknown Source) at org.apache.derby.impl.drda.DRDAConnThread.sendUnexpectedException(Unknown Source) at org.apache.derby.impl.drda.DRDAConnThread.handleException(Unknown Source) at org.apache.derby.impl.drda.DRDAConnThread.run(Unknown Source) used 1500 ms . **10) testFunctionNullHandling(org.apache.derbyTesting.functionTests.tests.lang.RoutineTest)junit.framework.AssertionFailedError: Column value mismatch @ column '1', row 1: Expected: >11:00:00< Found: >null< at org.apache.derbyTesting.junit.JDBC.assertRowInResultSet(JDBC.java:109... at org.apache.derbyTesting.functionTests.tests.lang.RoutineTest.testFunctionNullHandling(RoutineTest.java:575) Needs investigation. I don't see a specific change to the assertion since 10.5. *13) testUpdateXXXWithAllDatatypes(org.apache.derbyTesting.functionTests.tests.lang.UpdatableResultSetTest)junit.framework.AssertionFailedError: The expected SQL state must be five characters long caused by at org.apache.derbyTesting.functionTests.tests.lang.UpdatableResultSetTest.runTestUpdateXXXWithAllDatatypes(UpdatableResultSetTest.java:2726) Caused by: java.sql.SQLException: Stream or LOB value cannot be retrieved more than once The new error is expected but I am surprised that the SQLState would not be five characters. Needs Investigation. *22) Needs investigation. Test has not changed in years. testGreekCharacters(org.apache.derbyTesting.functionTests.tests.jdbcapi.DriverMgrAuthenticationTest)junit.framework.AssertionFailedError: Connection should've been refused/failed at org.apache.derbyTesting.functionTests.tests.jdbcapi.DriverMgrAuthenticationTest.assertConnectionFail(DriverMgrAuthenticationTest.java:116)
          Hide
          Kathey Marsden added a comment -

          Using the copied system/wombat database from 10_5SoftupgradeTestOutput.txt , I did a run (still in soft upgrade mode) with 10.5.3.2 - (1171883) derbyclient.jar and derbyTesting.jar (first in the classpath) and 10.8.2.1 for the other jars. I will attach as 0_5SoftUpgradeAnd10_5clientTestOutput.txt. There were Failures: 202, Errors: 72, most similar to the straight soft upgrade run.
          I have not done a full analysis yet, but filed DERBY-5429 for the incompatible JVMInfo.class in derbyclient.jar causing an issue triggering JVM dumps.

          The only protocol related error I saw was this error:
          119) testConnectShutdownAuthentication(org.apache.derbyTesting.functionTests.tests.jdbcapi.DriverMgrAuthenticationTest)junit.framework.ComparisonFailure: Unexpected SQL state. expected:<0800[4]> but was:<0800[6]>
          Caused by: org.apache.derby.client.am.DisconnectException: Insufficient data while reading from the network - expected a minimum of 6 bytes and received only 0 bytes. The connection has been terminated.

          I think that is also fallout from DERBY-5429 because the javaDump() method can't be found we get a serious error on the server side. causing a serious error server side which leads to the protocol issue.

          There were no other protocol issues or hangs which is mostly what I am looking for in this testing.

          Show
          Kathey Marsden added a comment - Using the copied system/wombat database from 10_5SoftupgradeTestOutput.txt , I did a run (still in soft upgrade mode) with 10.5.3.2 - (1171883) derbyclient.jar and derbyTesting.jar (first in the classpath) and 10.8.2.1 for the other jars. I will attach as 0_5SoftUpgradeAnd10_5clientTestOutput.txt. There were Failures: 202, Errors: 72, most similar to the straight soft upgrade run. I have not done a full analysis yet, but filed DERBY-5429 for the incompatible JVMInfo.class in derbyclient.jar causing an issue triggering JVM dumps. The only protocol related error I saw was this error: 119) testConnectShutdownAuthentication(org.apache.derbyTesting.functionTests.tests.jdbcapi.DriverMgrAuthenticationTest)junit.framework.ComparisonFailure: Unexpected SQL state. expected:<0800 [4] > but was:<0800 [6] > Caused by: org.apache.derby.client.am.DisconnectException: Insufficient data while reading from the network - expected a minimum of 6 bytes and received only 0 bytes. The connection has been terminated. I think that is also fallout from DERBY-5429 because the javaDump() method can't be found we get a serious error on the server side. causing a serious error server side which leads to the protocol issue. There were no other protocol issues or hangs which is mostly what I am looking for in this testing.
          Hide
          Myrna van Lunteren added a comment -

          Kathey, can this issue be closed now?

          Show
          Myrna van Lunteren added a comment - Kathey, can this issue be closed now?
          Hide
          Kathey Marsden added a comment -

          linking issues discovered during this testing.

          Show
          Kathey Marsden added a comment - linking issues discovered during this testing.
          Hide
          Kathey Marsden added a comment -

          Resolving this issue. I have done all the compat testing I will for 10.8, although it never seems like enough. The two major issues discovered have been resolved.
          Note: If mixed jars no longer work in 10.9 then the test infrastructure will need to be changed to allow isolated class loader for client.

          Show
          Kathey Marsden added a comment - Resolving this issue. I have done all the compat testing I will for 10.8, although it never seems like enough. The two major issues discovered have been resolved. Note: If mixed jars no longer work in 10.9 then the test infrastructure will need to be changed to allow isolated class loader for client.
          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:
              Kathey Marsden
              Reporter:
              Kathey Marsden
            • Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:

                Development