[
https://issues.apache.org/jira/browse/SPARK-21543?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16104549#comment-16104549
]
zhoukang commented on SPARK-21543:
----------------------------------
First i will describe my case:
There is a cluster with one node has much more resources,but this node has bad
disk.However, yarn will always launch container on this node even if executor
can not init successful on it.
Then job failed after 4 times retry.
I agree with you.This should consider together with blacklist.Since do not
count towards task failures will cause new problem.
I will close this issue and related pr.
I will think about optimize blacklist and add disk checker for shuffle server
register.
Thanks for your time [~srowen]
> Should not count executor initialize failed towards task failures
> -----------------------------------------------------------------
>
> Key: SPARK-21543
> URL: https://issues.apache.org/jira/browse/SPARK-21543
> Project: Spark
> Issue Type: Improvement
> Components: YARN
> Affects Versions: 1.6.1, 2.1.0
> Reporter: zhoukang
>
> Till now, when executor init failed and exit with error code = 1, it will
> count toward task failures.Which i think should not count executor initialize
> failed towards task failures.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
