[
https://issues.apache.org/jira/browse/FLINK-27921?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17558018#comment-17558018
]
Roman Khachatryan commented on FLINK-27921:
-------------------------------------------
There were some discussions in the past (e.g. in FLINK-25750), but nothing
specific AFAIK.
I think eventually we do need to solve this problem by somehow excluding job
setup time (the latter is or should already be measured by scheduler
benchmarks).
If we agree on that, I'd propose:
1. Create a separate ticket for the current regression
2. Update (configure) the affected or all benchmarks so they don't use the new
featur (and the results return back to normal)
3. Create a ticket to exclude job setup time
4. (gradually) update all the benchmarks (and remove the special config added
in (2))
WDYT?
> Introduce the checkResourceRequirementsWithDelay in DeclarativeSlotManager
> --------------------------------------------------------------------------
>
> Key: FLINK-27921
> URL: https://issues.apache.org/jira/browse/FLINK-27921
> Project: Flink
> Issue Type: Sub-task
> Components: Runtime / Coordination
> Reporter: Aitozi
> Assignee: Aitozi
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.16.0
>
>
> As discussed in
> [https://github.com/apache/flink/pull/19840#discussion_r884242067] .This
> ticket is meant to introduce the same mechanism to wait for a slight delay
> before process the resource check with {{{}FineGrainedSlotManager{}}}. It
> will reduce the frequency of unnecessary re-allocations
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
