[ https://issues.apache.org/jira/browse/FLINK-3764?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15242870#comment-15242870 ]
ASF GitHub Bot commented on FLINK-3764: --------------------------------------- Github user zentol commented on a diff in the pull request: https://github.com/apache/flink/pull/1894#discussion_r59864663 --- Diff: flink-runtime/src/test/scala/org/apache/flink/runtime/testingUtils/TestingCluster.scala --- @@ -211,11 +211,18 @@ class TestingCluster( val jmsAliveFutures = jobManagerActors map { _ map { - tm => (tm ? Alive)(timeout) + jm => (jm ? Alive)(timeout) + } + } getOrElse(Seq()) + + val resourceManagersAliveFutures = resourceManagerActors map { + _ map { + rm => (rm ? Alive)(timeout) } } getOrElse(Seq()) - val combinedFuture = Future.sequence(tmsAliveFutures ++ jmsAliveFutures) + val combinedFuture = Future.sequence(tmsAliveFutures ++ jmsAliveFutures ++ + resourceManagersAliveFutures) Await.ready(combinedFuture, timeout) --- End diff -- if this check fails, do we know due to which future? > LeaderChangeStateCleanupTest.testStateCleanupAfterListenerNotification fails > on Travis > -------------------------------------------------------------------------------------- > > Key: FLINK-3764 > URL: https://issues.apache.org/jira/browse/FLINK-3764 > Project: Flink > Issue Type: Bug > Reporter: Till Rohrmann > Assignee: Till Rohrmann > Priority: Critical > Labels: test-stability > > The > {{LeaderChangeStateCleanupTest.testStateCleanupAfterListenerNotification}} > fails spuriously on Travis because of a {{NullPointerException}}. The reason > is that it's not properly waited until the {{ResourceManager}} has been > started. Due to this, it can happen that a leader notification message is > tried to be sent to a {{LeaderRetrievalListener}} which has not been set by > the {{ResourceManager}}. > [1] https://s3.amazonaws.com/archive.travis-ci.org/jobs/123271732/log.txt -- This message was sent by Atlassian JIRA (v6.3.4#6332)