narendly commented on a change in pull request #690: Reset the WAGED rebalancer
once the controller newly acquires leadership.
URL: https://github.com/apache/helix/pull/690#discussion_r369862100
##########
File path:
helix-core/src/main/java/org/apache/helix/controller/GenericHelixController.java
##########
@@ -626,6 +631,22 @@ private void handleEvent(ClusterEvent event,
BaseControllerDataProvider dataProv
return;
}
+ // Event handling happens in a different thread from the
onControllerChange processing thread.
+ // Thus, there are several possible conditions.
+ // 1. Event handled after leadership acquired. So we will have a valid
rebalancer for the
+ // event processing.
+ // 2. Event handled shortly after leadership relinquished. And the
rebalancer has not been
+ // marked as invalid yet. So the event will be processed the same as case
one.
+ // 3. Event is leftover from the previous session, and it is handled when
the controller
+ // regains the leadership. The rebalancer will be reset before being used.
That is the
+ // expected behavior so as to avoid inconsistent rebalance result.
+ // 4. Event handled shortly after leadership relinquished. And the
rebalancer has been marked
+ // as invalid. So we reset the rebalancer. But the later isLeader() check
will return false and
Review comment:
Could you please explain why another pipeline is being triggered when the
controller is no longer a leader?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to 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
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
