You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@spark.apache.org by "Apache Spark (Jira)" <ji...@apache.org> on 2022/10/04 07:22:00 UTC

[jira] [Commented] (SPARK-40096) Finalize shuffle merge slow due to connection creation fails

    [ https://issues.apache.org/jira/browse/SPARK-40096?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17612497#comment-17612497 ] 

Apache Spark commented on SPARK-40096:
--------------------------------------

User 'wankunde' has created a pull request for this issue:
https://github.com/apache/spark/pull/38091

> Finalize shuffle merge slow due to connection creation fails
> ------------------------------------------------------------
>
>                 Key: SPARK-40096
>                 URL: https://issues.apache.org/jira/browse/SPARK-40096
>             Project: Spark
>          Issue Type: Bug
>          Components: Spark Core
>    Affects Versions: 3.3.0
>            Reporter: Wan Kun
>            Assignee: Wan Kun
>            Priority: Major
>             Fix For: 3.4.0
>
>
> *How to reproduce this issue*
>  * Enable push based shuffle
>  * Remove some merger nodes before sending finalize RPCs
>  * Driver try to connect those merger shuffle services and send finalize RPC one by one, each connection creation will timeout after SPARK_NETWORK_IO_CONNECTIONCREATIONTIMEOUT_KEY (120s by default)
>  
> We can send these RPCs in *shuffleMergeFinalizeScheduler*  thread pool and handle the connection creation exception



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

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org