[ 
https://issues.apache.org/jira/browse/SPARK-53898?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

wuyi updated SPARK-53898:
-------------------------
    Description: MapOutputTrackerMaster.shufflestatuses can be mistakenly 
cleaned by Shuffle Cleanup feature, leading to SparkException (crashing the 
SparkContext) by the subsequent access to that already removed shuffle 
metadata. A real case (limited to local cluster currently) is the ongoing 
subquery could access the shuffle metadata which has been already cleanedup 
after the main query completes. See the detailed discussion at: 
[https://github.com/apache/spark/pull/52213#discussion_r2415632474].  (was: 
MapOutputTrackerMaster.shufflestatuses can be mistakenly cleaned by Shuffle 
Cleanup feature, leading to SparkException (crashing the SparkContext) by the 
subsequent access to that already removed shuffle metadata. A real case is the 
ongoing subquery could access the shuffle metadata which has been already 
cleanedup after the main query completes. See the detailed discussion at: 
https://github.com/apache/spark/pull/52213#discussion_r2415632474.)

> MapOutputTrackerMaster.shufflestatuses is mistakenly cleaned by Shuffle 
> cleanup
> -------------------------------------------------------------------------------
>
>                 Key: SPARK-53898
>                 URL: https://issues.apache.org/jira/browse/SPARK-53898
>             Project: Spark
>          Issue Type: Bug
>          Components: Spark Core
>    Affects Versions: 4.0.0, 4.0.1
>            Reporter: wuyi
>            Priority: Major
>
> MapOutputTrackerMaster.shufflestatuses can be mistakenly cleaned by Shuffle 
> Cleanup feature, leading to SparkException (crashing the SparkContext) by the 
> subsequent access to that already removed shuffle metadata. A real case 
> (limited to local cluster currently) is the ongoing subquery could access the 
> shuffle metadata which has been already cleanedup after the main query 
> completes. See the detailed discussion at: 
> [https://github.com/apache/spark/pull/52213#discussion_r2415632474].



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to