Folks, I observe a lot of flaky tests in Apache Jenkins. They seem rather random and not tied to particular machines (saw failures on H1 and on H2). Moreover, there are no tickets for them and I haven't seen any of those failures in our internal CI.
Does anyone have an idea about any recent changes in test harness, libprocess or whatever that could lead to this? It's probably not related to MESOS-6180 <https://issues.apache.org/jira/browse/MESOS-6180>, because not all failures are future timeout induced. For example, in the last day I saw these guys failing: ReconciliationTest.RecoveredAgent [1] MasterTest.TaskLabels [2] RoleTest.ImplicitRoleRegister [3] ReconciliationTest.ImplicitTerminalTask [4] ReservationTest.BadACLDropReserve [5] ReservationTest.CompatibleCheckpointedResources [6] ContentType/SchedulerHttpApiTest.Subscribe/0 [7] [1] https://goo.gl/cs88BD [2] https://goo.gl/gTzKUV [3] https://goo.gl/7pGaQG [4] https://goo.gl/ccq38D [5] https://goo.gl/0R1eOO [6] https://goo.gl/xKQzUt [7] https://goo.gl/HZmiGJ