Description of problem: See Description of ISPN-3364
Vitalii Chepeliuk <vchepeli> made a comment on jira ISPN-3364 Added trace log for LocalDeltaAwarePassivationTest
Vitalii Chepeliuk <vchepeli> made a comment on jira ISPN-3364 Added trace log for ReplDeltaAwarePassivationTest
William Burns <wburns> updated the status of jira ISPN-3364 to Coding In Progress
William Burns <wburns> made a comment on jira ISPN-3364 This test does not fail on the latest ISPN master currently. I looked on Jenkins to try to find which commit caused the failure, but haven't had luck. I see a SHA-1 here at https://jenkins.mw.lab.eng.bos.redhat.com/hudson/view/JDG/view/FUNC/job/edg-60-ispn-testsuite-rhel/jdk=ibm17,label_exp=RHEL5_x86_64/166/parameters/ but it doesn't appear to be a valid ISPN commit SHA-1. Are you still able to reproduce this issue or can you point me to what commit this was used against?
Vitalii Chepeliuk <vchepeli> made a comment on jira ISPN-3364 Hi Will! You can still see it here. I run it with ER2, you can see it in matrix under IBM and JDK7, this tests fail every time https://jenkins.mw.lab.eng.bos.redhat.com/hudson/view/JDG/view/FUNC/job/edg-60-ispn-testsuite-rhel/
William Burns <wburns> updated the status of jira ISPN-3364 to Open
Vitalii Chepeliuk <vchepeli> made a comment on jira ISPN-3364 Hi Will! You can still see it here. I run it with ER3, you can see it in matrix under IBM7, this tests fail every time, here is build on RHEL6x86 machine https://jenkins.mw.lab.eng.bos.redhat.com/hudson/view/JDG/view/FUNC/job/edg-60-ispn-testsuite-rhel/USE_JDK=JDK7,jdk=ibm17,label_exp=RHEL6%20%26%26%20x86%20%26%26%20mem4G/lastCompletedBuild/testReport/ And RHEL6x86_64 machine https://jenkins.mw.lab.eng.bos.redhat.com/hudson/view/JDG/view/FUNC/job/edg-60-ispn-testsuite-rhel/USE_JDK=JDK7,jdk=ibm17,label_exp=RHEL6%20%26%26%20x86_64%20%26%26%20mem4G/lastCompletedBuild/testReport/
Vitalii Chepeliuk <vchepeli> made a comment on jira ISPN-3364 Hi Will! You can still see it here. I run it with ER3, you can see it in matrix under IBM7, this tests fail every time, here is build on RHEL6x86 machine https://jenkins.mw.lab.eng.bos.redhat.com/hudson/view/JDG/view/FUNC/job/edg-60-ispn-testsuite-rhel/USE_JDK=JDK7,jdk=ibm17,label_exp=RHEL6%20%26%26%20x86%20%26%26%20mem4G/227/#showFailuresLink And RHEL6x86_64 machine https://jenkins.mw.lab.eng.bos.redhat.com/hudson/view/JDG/view/FUNC/job/edg-60-ispn-testsuite-rhel/USE_JDK=JDK7,jdk=ibm17,label_exp=RHEL6%20%26%26%20x86_64%20%26%26%20mem4G/228/#showFailuresLink
William Burns <wburns> made a comment on jira ISPN-3364 I was finally able to reproduce this in a way that I could debug to see what is going on. And needless to say that hasn't even helped. In testing the ReplDeltaAwarePassivationTest I was able to narrow the culprit down to a single line where things were not working. In BoundedConcurrentHashMap:1719 it is finally assigning the new entry to the array. For some reason this affects both the DataContainer and the DummyInMemoryCacheStore entries maps and ends up removing the additional element. Looking at all the references between the dc and store there is nothing shared in the map and doesn't make much sense as to why this occurs. I have tried changing the hash maps that are used in both the DataContainer and DummyStore to use different implementations respectively but it hasn't change the outcome either. I haven't been able to reproduce the LocalDeltaAwarePassivationTest issue yet as it only is reproducible in a full maven run with all the tests.
William Burns <wburns> made a comment on jira ISPN-3364 I was finally able to reproduce this in a way that I could debug to see what is going on. And needless to say that hasn't even helped. In testing the ReplDeltaAwarePassivationTest I was able to narrow the culprit down to a single line where things were not working. In BoundedConcurrentHashMap:1719 it is finally assigning the new entry to the array. For some reason this affects both the DataContainer and the DummyInMemoryCacheStore entries maps and ends up removing the additional element. Looking at all the references between the dc and store there is nothing shared in the map and doesn't make much sense as to why this occurs. I have tried changing the hash maps that are used in both the DataContainer and DummyStore to use different implementations respectively but it hasn't change the outcome either. I haven't been able to debug fully the LocalDeltaAwarePassivationTest issue yet as it only is reproducible in a full maven run with all the tests.
William Burns <wburns> made a comment on jira ISPN-3364 Looking closer it appears that the changes in ISPN-1822 are now causing this issue. If I change it to use the standard LRU implementation all tests pass. My guess is that IBM finally fixed their issue with LinkedHashMap. I am wondering if it is worth keeping the old fixes or not for older IBM JVMs or not, it would be nice to just remove the code now.
Running tests with -Pjdk6testsuite profile and IBM6 See following report from last test run https://jenkins.mw.lab.eng.bos.redhat.com/hudson/view/JDG/view/FUNC/job/edg-60-ispn-testsuite-rhel/290/testReport/
Vitalii Chepeliuk <vchepeli> updated the status of jira ISPN-3364 to Reopened