[
https://issues.apache.org/jira/browse/SPARK-9823?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14682025#comment-14682025
]
Thomas Graves commented on SPARK-9823:
--------------------------------------
Sorry it turned out I had some really high timeouts configured. I tested this
again with default timeouts and it did exit properly.
> Spark application hangs when using dynamic allocations and executor fails
> connect to shuffle service
> ----------------------------------------------------------------------------------------------------
>
> Key: SPARK-9823
> URL: https://issues.apache.org/jira/browse/SPARK-9823
> Project: Spark
> Issue Type: Bug
> Components: YARN
> Affects Versions: 1.5.0
> Reporter: Thomas Graves
> Priority: Critical
>
> I was trying out spark 1.5 with the dynamic allocations turned on in YARN.
> If the executor fails to connect to the external shuffle service then the
> application just hangs there. The SparkContext didn't seem to be doing
> anything or requesting any new executors to replace it.
> In this case the shuffle service had security on but I forgot to turn it on
> in the application.
> ====
> This is somewhat a tangent but the reason I tried spark 1.5 is with
> spark1.4.1 I also found the application hung with no executors so it must not
> have been requesting them properly.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
