zentol commented on code in PR #22169:
URL: https://github.com/apache/flink/pull/22169#discussion_r1147328427
##########
flink-runtime/src/main/java/org/apache/flink/runtime/scheduler/adaptive/Executing.java:
##########
@@ -123,9 +123,18 @@ private void handleDeploymentFailure(ExecutionVertex
executionVertex, JobExcepti
}
@Override
- public void notifyNewResourcesAvailable() {
- if (context.canScaleUp(getExecutionGraph())) {
- getLogger().info("New resources are available. Restarting job to
scale up.");
+ public void onNewResourcesAvailable() {
+ maybeRescale();
+ }
+
+ @Override
+ public void onNewResourceRequirements() {
+ maybeRescale();
+ }
Review Comment:
> do you mean it contradicts the current behavior of Reactive Scheduling?
Reactive mode isn't affected because there the actual requirements don't
change.
> I think there's no contradiction with "scale down whenever possible" if we
interpret it as "scale down whenever cumulative parallelism decreases".
That would indeed not be contradictory, but not really a desirable behavior?
Not without some fallback condition at least... (like a timeout).
> Shall we create a ticket for that? (maybe one catch-all ticket for now)
Will do.
--
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.
To unsubscribe, e-mail: issues-unsubscr...@flink.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
