GJL commented on a change in pull request #6451: [FLINK-9936] Resource manager
connect to mesos after leadership granted. .
URL: https://github.com/apache/flink/pull/6451#discussion_r206522513
##########
File path:
flink-mesos/src/main/java/org/apache/flink/mesos/runtime/clusterframework/MesosResourceManager.java
##########
@@ -221,10 +221,14 @@ protected ActorRef
createReconciliationCoordinator(SchedulerDriver schedulerDriv
//
------------------------------------------------------------------------
/**
- * Starts the Mesos-specifics.
+ * Do nothing and all work has been moved to on leadership granted
callback.
*/
@Override
protected void initialize() throws ResourceManagerException {
+ }
+
+ @Override
+ protected void onLeaderShipGranted() throws Exception {
Review comment:
We are re-creating the supporting actors without shutting them down. This
should result in memory leaks everytime the instance wins the leader election.
It is also not very nice that we occupy the main thread of the actor with
blocking IO.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
