[
https://issues.apache.org/jira/browse/SPARK-60570?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Wenchen Fan resolved SPARK-60570.
---------------------------------
Fix Version/s: 4.4.0
Resolution: Fixed
Issue resolved by pull request 59801
[https://github.com/apache/spark/pull/59801]
> Add AccumulatorContext.getOrNull to avoid an Option per driver-side
> accumulator lookup
> --------------------------------------------------------------------------------------
>
> Key: SPARK-60570
> URL: https://issues.apache.org/jira/browse/SPARK-60570
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
> Affects Versions: 4.4.0
> Reporter: Shujing Yang
> Assignee: Shujing Yang
> Priority: Major
> Labels: pull-request-available
> Fix For: 4.4.0
>
>
> DAGScheduler.updateAccumulators runs on the driver for every finished task
> and looks up every accumulator update the task carries, which can be hundreds
> when SQL metrics are involved. AccumulatorContext.get returns an Option, so
> each lookup allocates a Some that is dropped right away. Add
> AccumulatorContext.getOrNull, which does the same lookup but returns null
> when the accumulator is not registered or has been garbage collected. get
> delegates to it, and updateAccumulators uses it directly. No behavior change.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
