1 2 Previous Next 17 Replies Latest reply on May 22, 2008 2:54 AM by timfox

    State of our failing tests

    jmesnil

      As we race towards JBM2 alpha, I did some investigation on our current failing tests.
      Some are failing because of missing features (ping, message chunk) or identified issues (security checks, unicod encodings).
      The rest of the failing bugs seems related to acknowledgement & redelivery issues

      Related to disabled ping mechanism:

      org.jboss.messaging.tests.integration.core.remoting.impl.ClientCrashTest.testCrashClientWithOneConnection
      org.jboss.messaging.tests.integration.core.remoting.impl.ClientCrashTest.testCrashClientWithTwoConnections
      org.jboss.messaging.tests.integration.core.remoting.mina.ClientKeepAliveTest.testKeepAliveWithClientOK
      org.jboss.messaging.tests.integration.core.remoting.mina.ClientKeepAliveTest.testKeepAliveWithClientNotResponding
      org.jboss.messaging.tests.integration.core.remoting.mina.ClientKeepAliveTest.testKeepAliveWithClientTooLongToRespond
      org.jboss.messaging.tests.integration.core.remoting.mina.ClientKeepAliveTest.testKeepAliveWithClientRespondingAndClientNotResponding
      org.jboss.messaging.tests.integration.core.remoting.mina.ServerKeepAliveTest.testKeepAliveWithServerNotResponding
      org.jboss.messaging.tests.integration.core.remoting.mina.ServerKeepAliveTest.testKeepAliveWithServerSessionFailed
      org.jboss.messaging.tests.unit.jms.network.ClientNetworkFailureTest.testServerResourcesCleanUpWhenClientCommThrowsException
      org.jboss.messaging.tests.unit.jms.network.ClientNetworkFailureTest.testServerResourcesCleanUpWhenClientCommDropsPacket

      Related to message chunk:

      org.jboss.test.messaging.jms.String64KLimitTest.test64KLimitWithTextMessage
      org.jboss.test.messaging.jms.String64KLimitTest.test64KLimitWithObjectMessage

      Related to security check on one-way messages:

      org.jboss.test.messaging.jms.SecurityTest.testInvalidTopicPublisher
      org.jboss.test.messaging.jms.SecurityTest.testInvalidQueueSender
      org.jboss.test.messaging.jms.SecurityTest.testDefaultSecurityInvalid

      Related to Unicode encodings:

      org.jboss.test.messaging.jms.message.MessageBodyTest.testMapMessage
      org.jboss.messaging.tests.unit.core.util.SimpleStringTest.testUnicode

      Related to Journal

      org.jboss.messaging.tests.unit.core.journal.impl.FakeJournalImplTest.testReclaimTransactionalAddCommit
      org.jboss.messaging.tests.unit.core.journal.impl.FakeJournalImplTest.testCommitRecordsInFileNoReclaim
      org.jboss.messaging.tests.unit.core.journal.impl.FakeJournalImplTest.testPrepareNoReclaim
      org.jboss.messaging.tests.unit.core.journal.impl.FakeJournalImplTest.testRollbackRecordsInFileNoReclaim
      org.jboss.messaging.tests.unit.core.journal.impl.FakeJournalImplTest.testPrepareReclaim

      Not investigated

      org.jboss.test.messaging.jms.MessageConsumerTest.testRedelMessageListener2
      org.jboss.test.messaging.jms.message.BytesMessageTest.testRedelivery
      org.jboss.test.messaging.jms.AcknowledgementTest.testMessageListenerClientAck
      org.jboss.test.messaging.jms.AcknowledgementTest.testMessageListenerTransactionalAck
      org.jboss.test.messaging.jms.AcknowledgementTest.testAutoAcknowledge
      org.jboss.test.messaging.jms.AcknowledgementTest.testDupsOKAcknowledgeQueue
      org.jboss.test.messaging.jms.AcknowledgementTest.testMessageListenerAutoAck
      org.jboss.test.messaging.jms.AcknowledgementTest.testMessageListenerDupsOK
      org.jboss.test.messaging.jms.CTSMiscellaneousTest.testNonPersistentMessagesSentSynchronously
      org.jboss.test.messaging.jms.MessageConsumerTest.testExceptionMessageListener1
      org.jboss.test.messaging.jms.MessageConsumerTest.testExceptionMessageListener2

        • 1. Re: State of our failing tests
          ataylor

           

          Related to security check on one-way messages:

          org.jboss.test.messaging.jms.SecurityTest.testInvalidTopicPublisher
          org.jboss.test.messaging.jms.SecurityTest.testInvalidQueueSender
          org.jboss.test.messaging.jms.SecurityTest.testDefaultSecurityInvalid


          my next commit will fix these

          • 2. Re: State of our failing tests
            ataylor

             

            Related to disabled ping mechanism:

            org.jboss.messaging.tests.integration.core.remoting.impl.ClientCrashTest.testCrashClientWithOneConnection
            org.jboss.messaging.tests.integration.core.remoting.impl.ClientCrashTest.testCrashClientWithTwoConnections
            org.jboss.messaging.tests.integration.core.remoting.mina.ClientKeepAliveTest.testKeepAliveWithClientOK
            org.jboss.messaging.tests.integration.core.remoting.mina.ClientKeepAliveTest.testKeepAliveWithClientNotResponding
            org.jboss.messaging.tests.integration.core.remoting.mina.ClientKeepAliveTest.testKeepAliveWithClientTooLongToRespond
            org.jboss.messaging.tests.integration.core.remoting.mina.ClientKeepAliveTest.testKeepAliveWithClientRespondingAndClientNotResponding
            org.jboss.messaging.tests.integration.core.remoting.mina.ServerKeepAliveTest.testKeepAliveWithServerNotResponding
            org.jboss.messaging.tests.integration.core.remoting.mina.ServerKeepAliveTest.testKeepAliveWithServerSessionFailed
            org.jboss.messaging.tests.unit.jms.network.ClientNetworkFailureTest.testServerResourcesCleanUpWhenClientCommThrowsException
            org.jboss.messaging.tests.unit.jms.network.ClientNetworkFailureTest.testServerResourcesCleanUpWhenClientCommDropsPacket


            my next but one commit will fix these

            • 3. Re: State of our failing tests
              ataylor

               

              Related to message chunk:

              org.jboss.test.messaging.jms.String64KLimitTest.test64KLimitWithTextMessage
              org.jboss.test.messaging.jms.String64KLimitTest.test64KLimitWithObjectMessage


              If chunking is going into Beta then we should comment out and raise a JIRA

              • 4. Re: State of our failing tests
                timfox

                I'll take a look at the journal tests if no-one objects.

                • 5. Re: State of our failing tests
                  timfox

                   

                  "timfox" wrote:
                  I'll take a look at the journal tests if no-one objects.


                  These are failing due to changes Clebert made for the aio journal.

                  I'll wait for him to wake up :)

                  • 6. Re: State of our failing tests
                    timfox

                    Ack tests are failing since we send acks default async now. Just need to set block-on-acknowledge to true in jbm-configuration.xml for the tests only.

                    Andy is handling this.

                    • 7. Re: State of our failing tests
                      timfox

                       

                      "timfox" wrote:
                      Ack tests are failing since we send acks default async now. Just need to set block-on-acknowledge to true in jbm-configuration.xml for the tests only.

                      Andy is handling this.


                      Actually the param is in the connection factory config

                      • 8. Re: State of our failing tests
                        timfox

                        The unicode encoding issue on SimpleString is now fixed

                        • 9. Re: State of our failing tests
                          timfox

                          For org.jboss.test.messaging.jms.CTSMiscellaneousTest.testNonPersistentMessagesSentSynchronously

                          I suggest we have a flag "sendNonPersistentMessagesSynchronously" which defaults to false.

                          The test suite can override this to true;

                          • 10. Re: State of our failing tests
                            timfox

                            org.jboss.test.messaging.jms.CTSMiscellaneousTest.testNonPersistentMessagesSentSynchronously is fixed

                            • 11. Re: State of our failing tests
                              ataylor

                              security tests fixed.

                              half of ack tests fixed

                              • 12. Re: State of our failing tests
                                timfox

                                I also get the following failures in the aio journal tests:

                                testCommitRecordsInFileNoReclaim Failure expected:<1> but was:<0>
                                junit.framework.AssertionFailedError: expected:<1> but was:<0>
                                at org.jboss.messaging.tests.unit.core.journal.impl.JournalImplTestUnit.testCommitRecordsInFileNoReclaim(JournalImplTestUnit.java:1319)

                                0.021
                                testRollbackRecordsInFileNoReclaim Failure expected:<1> but was:<0>
                                junit.framework.AssertionFailedError: expected:<1> but was:<0>
                                at org.jboss.messaging.tests.unit.core.journal.impl.JournalImplTestUnit.testRollbackRecordsInFileNoReclaim(JournalImplTestUnit.java:1416)

                                0.027
                                testPrepareNoReclaim Failure expected:<1> but was:<0>
                                junit.framework.AssertionFailedError: expected:<1> but was:<0>
                                at org.jboss.messaging.tests.unit.core.journal.impl.JournalImplTestUnit.testPrepareNoReclaim(JournalImplTestUnit.java:1523)

                                0.022
                                testPrepareReclaim

                                • 13. Re: State of our failing tests
                                  clebert.suconic

                                  As the close is done asynchronous... I need to change the test to wait the completion of closes and callbacks for the test to be valid again.

                                  I have added a method on TestableJournal (debugWait).

                                  Adding the call to the wait on the test fixed the problem.


                                  I have changed it locally and it is fixed. I will commit it shortly.

                                  • 14. Re: State of our failing tests
                                    clebert.suconic

                                    The journal tests should be fixed now.
                                    I have run them few times.


                                    if anyone sees any failures please let me know as it will be some intermittent failure that needs to be investigated.

                                    1 2 Previous Next