spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Hyukjin Kwon (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (SPARK-15336) Detect and take action about executors running on corrupted nodes
Date Tue, 21 May 2019 04:25:18 GMT

     [ https://issues.apache.org/jira/browse/SPARK-15336?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Hyukjin Kwon updated SPARK-15336:
---------------------------------
    Labels: bulk-closed  (was: )

> Detect and take action about executors running on corrupted nodes
> -----------------------------------------------------------------
>
>                 Key: SPARK-15336
>                 URL: https://issues.apache.org/jira/browse/SPARK-15336
>             Project: Spark
>          Issue Type: Improvement
>            Reporter: He Tianyi
>            Priority: Minor
>              Labels: bulk-closed
>
> In a cluster consists of many nodes, there is a bigger chance that some corrupted nodes
are running and launching containers on it. Corruptions may due to OS configuration, or file
system (i.e. `python` binary corrupted in `pyspark`).
> In this case, if there are one executor running on one of these nodes, one may experience
a repeated task failure for tasks scheduled on that executor. 
> In production, this may also leads to job failure.
> Shall we provide a mechanism to detect it and do something like killing the executor?



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

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


Mime
View raw message