[
https://issues.apache.org/jira/browse/SPARK-16573?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15379862#comment-15379862
]
Sean Owen commented on SPARK-16573:
-----------------------------------
I am not so sure this is a Spark issue per se. You can accomplish this by
querying to the stderrs via the web API, or maybe by using YARN's log
aggregation. There are already ways to do it.
> executor stderr processing tools
> ---------------------------------
>
> Key: SPARK-16573
> URL: https://issues.apache.org/jira/browse/SPARK-16573
> Project: Spark
> Issue Type: Improvement
> Components: Input/Output
> Environment: spark 1.6.0
> Reporter: Norman He
> Original Estimate: 672h
> Remaining Estimate: 672h
>
> 1) from eacch executor, I can view stderr of each one
> 2) I would like to filter all the stderr of every executor ( saying I am
> interested #18 output of all executor ) and I would like to add the number in
> them by grepping certain field and sum them up.
> 3) TensorBoard having this kind of capability
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
