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 2021/04/01 09:26:00 UTC

[jira] [Commented] (SPARK-34934) Race condition while registering source in MetricsSystem

    [ https://issues.apache.org/jira/browse/SPARK-34934?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17313044#comment-17313044 ] 

Apache Spark commented on SPARK-34934:
--------------------------------------

User 'BOOTMGR' has created a pull request for this issue:
https://github.com/apache/spark/pull/32024

> Race condition while registering source in MetricsSystem 
> ---------------------------------------------------------
>
>                 Key: SPARK-34934
>                 URL: https://issues.apache.org/jira/browse/SPARK-34934
>             Project: Spark
>          Issue Type: Bug
>          Components: Spark Core
>    Affects Versions: 2.3.2, 3.1.1
>            Reporter: Harsh Panchal
>            Priority: Minor
>
> {{MetricsSystem}} manages {{mutable.ArrayBuffer}} of metric sources. {{registerSource}} and {{removeSource}} methods are provided to add/remove new source from Metric system. Both these methods are not synchronised. Also, underlying {{mutable.ArrayBuffer}} not being thread safe, unexpected behaviours are possible if called concurrently (in our case, multiple executors registering custom source).
> For ex, we got below exception due to race condition
> {noformat}
> java.lang.ArrayIndexOutOfBoundsException
> 	at scala.collection.mutable.ResizableArray$class.ensureSize(ResizableArray.scala:104)
> 	at scala.collection.mutable.ArrayBuffer.ensureSize(ArrayBuffer.scala:48)
> 	at scala.collection.mutable.ArrayBuffer.$plus$eq(ArrayBuffer.scala:84)
> 	at org.apache.spark.metrics.MetricsSystem.registerSource(MetricsSystem.scala:157)
> {noformat}



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

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