echauchot commented on code in PR #22985:
URL: https://github.com/apache/flink/pull/22985#discussion_r1360383883
##########
flink-runtime/src/main/java/org/apache/flink/runtime/scheduler/adaptive/Executing.java:
##########
@@ -67,13 +77,36 @@ class Executing extends StateWithExecutionGraph implements
ResourceListener {
this.context = context;
Preconditions.checkState(
executionGraph.getState() == JobStatus.RUNNING, "Assuming
running execution graph");
+ this.scalingIntervalMin = scalingIntervalMin;
+ this.scalingIntervalMax = scalingIntervalMax;
+ this.lastRescale =
+ Instant.now(); // Executing is recreated with each restart
(when we rescale)
+ // we consider the first execution of the pipeline as a rescale event
+ Preconditions.checkState(
+ !scalingIntervalMin.isNegative(),
+ "%s must be positive integer or 0",
+ JobManagerOptions.SCHEDULER_SCALING_INTERVAL_MIN.key());
+ if (scalingIntervalMax != null) {
+ Preconditions.checkState(
+ scalingIntervalMax.compareTo(scalingIntervalMin) > 0,
+ "%s(%d) must be greater than %s(%d)",
+ JobManagerOptions.SCHEDULER_SCALING_INTERVAL_MAX.key(),
+ scalingIntervalMax,
+ JobManagerOptions.SCHEDULER_SCALING_INTERVAL_MIN.key(),
+ scalingIntervalMin);
+ }
deploy();
// check if new resources have come available in the meantime
context.runIfState(this, this::maybeRescale, Duration.ZERO);
}
+ @VisibleForTesting
+ void setLastRescale(Instant lastRescale) {
Review Comment:
Yes I know this and I agree but the big difference to me is that this field
is a **constant** in the production code. But fair enough, I'll do the change
if you want to.
--
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
