[ https://issues.apache.org/jira/browse/FLINK-12895?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16869017#comment-16869017 ]
Till Rohrmann commented on FLINK-12895: --------------------------------------- I looked into the problem and I think that the original problem has been caused by FLINK-11059 and should be fixed with FLINK-12863. The reason why the test case also failed with FLINK-12863 was because I tried to harden the test case {{TaskExecutorTest#testSyncSlotsWithJobMasterByHeartbeat}} by directly executing the response from {{JobMasterGateway#offerSlots}} in the {{TaskExecutor}}. This was the wrong approach since it changed the order between heartbeats and slot offering response. > TaskManagerProcessFailureBatchRecoveryITCase.testTaskManagerProcessFailure > failed on travis > -------------------------------------------------------------------------------------------- > > Key: FLINK-12895 > URL: https://issues.apache.org/jira/browse/FLINK-12895 > Project: Flink > Issue Type: Bug > Components: Runtime / Coordination, Tests > Affects Versions: 1.9.0 > Reporter: Haibo Sun > Priority: Critical > Labels: test-stability > > Logs: [https://api.travis-ci.org/v3/job/547509708/log.txt] > Build: [https://travis-ci.org/apache/flink/builds/547509701] -- This message was sent by Atlassian JIRA (v7.6.3#76005)