I know that Till observed issues with the ScalaShell tests as well. If I remember correctly, the issues were caused by checking the log file for specific order of events, which is not necessarily the case.
– Ufuk > On 13 Oct 2015, at 16:14, Sachin Goel <sachingoel0...@gmail.com> wrote: > > Hi all > Over the past few days, I've observed multiple failures of this test case. > There is no single reason. > 1. The assertion to check the output fails [either the finished status or > the actual job output]. > https://travis-ci.org/apache/flink/jobs/84664370 > 2. JVM runs out of heap space [weirdly]. > 3. Just recently, observed a loss of the task manager. > https://travis-ci.org/apache/flink/jobs/85108827 > > Has someone else observed any such issues? It seems a simple enough test to > display such flaky behavior. > > Cheers! > Sachin > -- Sachin Goel > Computer Science, IIT Delhi > m. +91-9871457685
