[
https://issues.apache.org/jira/browse/FLINK-35035?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17836395#comment-17836395
]
yuanfenghu edited comment on FLINK-35035 at 4/12/24 6:08 AM:
-------------------------------------------------------------
[~echauchot]
Thank you for your reply.
I think you are looking at this scene from the perspective of Reactive Mode,
because Reactive Mode only uses the resources of the cluster as a criterion for
task parallelism. I don’t know if I understand it correctly.
But my above scenario is in non-Reactive Mode. I just use the adaptive
scheduler, which means that I increase the parallelism of the running task from
10 to 12. However, because min-parallelism-increase=5, I am satisfied in the
cluster slot. When the condition of 12 is met, the expansion of the task cannot
be triggered immediately, but it needs to wait for scaling-interval.max before
the expansion can be triggered. My purpose is to trigger the expansion when the
parallelism of 12 is met, instead of having to after scaling-interval.max or
min-parallelism-increase
was (Author: JIRAUSER296932):
[~echauchot]
Thank you for your reply.
You should look at this issue from the perspective of Reactive Mode, because
Reactive Mode only uses the resources of the cluster as a criterion for task
parallelism. I don’t know if I understand it correctly.
But my above scenario is in non-Reactive Mode. But I use the adaptive
scheduler, which means that I increase the parallelism of the running task from
10 to 12. However, because min-parallelism-increase=5, I am satisfied in the
cluster slot. When the condition of 12 is met, the expansion of the task cannot
be triggered immediately, but it needs to wait for scaling-interval.max before
the expansion can be triggered. My purpose is to trigger the expansion when the
parallelism of 12 is met, instead of having to after scaling-interval.max
> Reduce job pause time when cluster resources are expanded in adaptive mode
> --------------------------------------------------------------------------
>
> Key: FLINK-35035
> URL: https://issues.apache.org/jira/browse/FLINK-35035
> Project: Flink
> Issue Type: Improvement
> Components: Runtime / Task
> Affects Versions: 1.19.0
> Reporter: yuanfenghu
> Priority: Minor
>
> When 'jobmanager.scheduler = adaptive' , job graph changes triggered by
> cluster expansion will cause long-term task stagnation. We should reduce this
> impact.
> As an example:
> I have jobgraph for : [v1 (maxp=10 minp = 1)] -> [v2 (maxp=10, minp=1)]
> When my cluster has 5 slots, the job will be executed as [v1 p5]->[v2 p5]
> When I add slots the task will trigger jobgraph changes，by
> org.apache.flink.runtime.scheduler.adaptive.ResourceListener#onNewResourcesAvailable，
> However, the five new slots I added were not discovered at the same time (for
> convenience, I assume that a taskmanager has one slot), because no matter
> what environment we add, we cannot guarantee that the new slots will be added
> at once, so this will cause onNewResourcesAvailable triggers repeatedly
> ，If each new slot action has a certain interval, then the jobgraph will
> continue to change during this period. What I hope is that there will be a
> stable time to configure the cluster resources and then go to it after the
> number of cluster slots has been stable for a certain period of time. Trigger
> jobgraph changes to avoid this situation
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
