Maxim Khutornenko created AURORA-1046:
-----------------------------------------
Summary: Memory leak in CachedClusterState
Key: AURORA-1046
URL: https://issues.apache.org/jira/browse/AURORA-1046
Project: Aurora
Issue Type: Bug
Components: Scheduler
Reporter: Maxim Khutornenko
This happens due to StateManager.assignTask() implementation [1] firing a
TaskStateChange event before updating the assignedTask slaveId and slaveHost
fields. As a result, all ASSIGNED preemption victims fall under a null key [2]
thus leaking memory over time.
As a side effect, it makes it impossible for the preemptor to preempt tasks in
ASSIGNED state.
[1] -
https://github.com/apache/incubator-aurora/blob/master/src/main/java/org/apache/aurora/scheduler/state/StateManagerImpl.java#L169-L212
[2] -
https://github.com/apache/incubator-aurora/blob/master/src/main/java/org/apache/aurora/scheduler/async/preemptor/CachedClusterState.java#L44
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
