[
https://issues.apache.org/jira/browse/YUNIKORN-3351?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ASF GitHub Bot updated YUNIKORN-3351:
-------------------------------------
Labels: pull-request-available (was: )
> Hoist the per-ask backoff getter out of tryAllocate
> ---------------------------------------------------
>
> Key: YUNIKORN-3351
> URL: https://issues.apache.org/jira/browse/YUNIKORN-3351
> Project: Apache YuniKorn
> Issue Type: Sub-task
> Components: core - scheduler
> Reporter: Dale Richardson
> Assignee: Dale Richardson
> Priority: Major
> Labels: pull-request-available
>
> {{Application.tryAllocate}} calls {{sa.queue.GetMaxAppUnschedAskBackoff()}}
> inside the loop that walks the application's asks, so a value that is
> constant for the whole scheduling cycle is re-fetched ( under the queue lock)
> once per ask walked.
> Hoist it above the loop. The entry guard changes from {{sa.sortedRequests ==
> nil}} to {{len(sa.sortedRequests) == 0}} so the hoisted call still only
> happens when there is at least one ask to walk, preserving the existing "not
> called when there is no work" property.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
