While we should definitely eventually address the failures Attila saw with Hadoop 2.6.0; I think the minimum for an Oozie 4.3.0 release should be to have no test failures with Hadoop 2.4.0 (the default). We tend to have a lot of flakey tests that we ignore during development, and this might be a good impetus to force ourselves to fix them.
- Robert On Fri, Oct 28, 2016 at 4:29 PM, satish saley <satishsale...@gmail.com> wrote: > Hi, > > I see three test failures, will create a jira. > > > testForkJoinMismatch(org.apache.oozie.workflow.lite. > TestLiteWorkflowAppParser) > [OOZIE-2719] > > > testActionKillCommandDate(org.apache.oozie.command.coord. > TestCoordActionsKillXCommand): > expected:<RUNNING> but was:<KILLED> > > > testMaxMatThrottleNotPicked(org.apache.oozie.service. > TestCoordMaterializeTriggerService): > expected:<Fri Oct 28 13:33:38 PDT 2016> but was:<Fri Oct 28 13:33:40 PDT > 2016> > > On Fri, Oct 28, 2016 at 3:26 AM, Attila Sasvari <asasv...@cloudera.com> > wrote: > > > Update: JavaActionExecutor tests seem to always fail with hadoop 2.6.0 , > > with 2.4.0 these tests pass. > > > > On Fri, Oct 28, 2016 at 11:22 AM, Attila Sasvari <asasv...@cloudera.com> > > wrote: > > > > > Hi there, > > > > > > I also downloaded it and tried to build it with running the > > > JavaActionExecutor tests: > > > > > > mvn clean package assembly:single -DjavaVersion=1.8 -DtargetVersion=1.7 > > > -Dtest=org.apache.oozie.action.hadoop.TestJavaActionExecutor > -Phadoop-2 > > > -Dhadoop.version=*2.6.0* -Puber > > > > > > > > > Failed tests: testSimpestSleSubmitOK(org.apa > > > che.oozie.action.hadoop.TestJavaActionExecutor) > > > testOutputSubmitOK(org.apache.oozie.action.hadoop.TestJavaAc > > > tionExecutor) > > > testIdSwapSubmitOK(org.apache.oozie.action.hadoop.TestJavaAc > > > tionExecutor) > > > testAdditionalJarSubmitOK(org.apache.oozie.action.hadoop.Tes > > > tJavaActionExecutor) > > > testExit0SubmitOK(org.apache.oozie.action.hadoop. > > TestJavaActionExecutor) > > > testExit1SubmitError(org.apache.oozie.action.hadoop.TestJava > > > ActionExecutor) > > > testChildKill(org.apache.oozie.action.hadoop. > TestJavaActionExecutor): > > > expected:<1> but was:<0> > > > testExceptionSubmitException(org.apache.oozie.action.hadoop. > > > TestJavaActionExecutor) > > > testExceptionSubmitThrowable(org.apache.oozie.action.hadoop. > > > TestJavaActionExecutor) > > > testRecovery(org.apache.oozie.action.hadoop.TestJavaActionExecutor): > > > expected:<...ob_1477579833383_001[1]> but > was:<...ob_1477579833383_001[ > > 2]> > > > testPrepare(org.apache.oozie.action.hadoop.TestJavaActionExecutor) > > > testUpdateConfForTimeLineServiceEnabled(org.apache.oozie.act > > > ion.hadoop.TestJavaActionExecutor): expected:<[tru]e> but > was:<[fals]e> > > > testActionShareLibWithNonDefaultNamenode(org.apache.oozie.ac > > > tion.hadoop.TestJavaActionExecutor) > > > testJobSubmissionWithoutYarnKill(org.apache.oozie.action.had > > > oop.TestJavaActionExecutor) > > > > > > Tests run: 46, Failures: 14, Errors: 0, Skipped: 0 > > > > > > (using OS X 10.11.6 (15G1004), java 1.8.0_101) > > > > > > Regards, > > > - Attila > > > > > > ps: I am re-running the tests to verify it is not a sporadic problem > that > > > is specific to my local dev environment. > > > > > > > > > On Tue, Oct 25, 2016 at 7:38 AM, Abhishek Bafna <bafna.i...@gmail.com> > > > wrote: > > > > > >> Hi, > > >> > > >> I have created a build for Oozie 4.3.0, candidate 0. > > >> > > >> Keys to verify the signature of the release artifact are available at > > >> > > >> http://www.apache.org/dist/oozie/KEYS > > >> > > >> Please download, test, and try it out: > > >> > > >> https://dist.apache.org/repos/dist/dev/oozie/oozie-4.3.0-rc0/ > > >> > > >> The release, md5 signature, gpg signature, and rat report can all > > >> be found at the above address. > > >> > > >> Vote closes on [Sat Oct 29 11:00:00 IST 2016]. > > >> > > >> Thanks! > > >> Abhishek Bafna > > >> > > > > > > > > >