[ https://issues.apache.org/jira/browse/MESOS-1013?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14694199#comment-14694199 ]
Greg Mann commented on MESOS-1013: ---------------------------------- Similar to MESOS-830, this test has been flaky because the {{ZooKeeperTestServer}} shutdown may not be complete before {{System.exit()}} is called, leading to native destructors being called on glog mutexes that are still in use. Unfortunately, there is no easy fix for this, and I've settled for a brief sleep before {{System.exit()}}. Review here: https://reviews.apache.org/r/37414/ > ExamplesTest.JavaLog is flaky > ----------------------------- > > Key: MESOS-1013 > URL: https://issues.apache.org/jira/browse/MESOS-1013 > Project: Mesos > Issue Type: Bug > Components: test > Affects Versions: 0.19.0 > Reporter: Vinod Kone > Assignee: Greg Mann > Labels: flaky, mesosphere > Attachments: ExamplesTest.JavaLog.logs > > > The {{ExamplesTest.JavaLog}} test framework is flaky, possibly related to a > race condition between mutexes. > {noformat} > [ RUN ] ExamplesTest.JavaLog > Using temporary directory '/tmp/ExamplesTest_JavaLog_WBWEb9' > Feb 18, 2014 12:10:57 PM TestLog main > INFO: Starting a local ZooKeeper server > ... > F0218 12:10:58.575036 17450 coordinator.cpp:394] Check failed: !missing Not > expecting local replica to be missing position 3 after the writing is done > *** Check failure stack trace: *** > tests/script.cpp:81: Failure > Failed > java_log_test.sh terminated with signal 'Aborted' > [ FAILED ] ExamplesTest.JavaLog (2166 ms) > {noformat} > Full logs attached. -- This message was sent by Atlassian JIRA (v6.3.4#6332)