spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sean Owen (JIRA)" <j...@apache.org>
Subject [jira] [Resolved] (SPARK-13133) When the option --master of spark-submit script is inconsistent with SparkConf.setMaster in Spark appliction code, the behavior of Spark application is difficult to understand
Date Mon, 15 Feb 2016 09:15:18 GMT

     [ https://issues.apache.org/jira/browse/SPARK-13133?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Sean Owen resolved SPARK-13133.
-------------------------------
    Resolution: Not A Problem

[~l90006322] don't reopen an issue unless the discussion has materially changed. I am quite
sure this is an application error. You set master twice; don't do that. 

This is not how you submit patches anyway.

> When the option --master of spark-submit script is inconsistent with SparkConf.setMaster
in Spark appliction code, the behavior of Spark application is difficult to understand
> -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
>                 Key: SPARK-13133
>                 URL: https://issues.apache.org/jira/browse/SPARK-13133
>             Project: Spark
>          Issue Type: Bug
>          Components: Deploy
>    Affects Versions: 1.3.0, 1.4.0, 1.5.0, 1.6.0
>            Reporter: Li Ye
>            Priority: Minor
>         Attachments: SPARK-13133-1.patch
>
>   Original Estimate: 168h
>  Remaining Estimate: 168h
>
> When the option --master of spark-submit script is inconsistent with SparkConf.setMaster
in Spark application code, the behavior is difficult to understand. For example, if the option
--master of spark-submit script is yarn-cluster while there is SparkConf.setMaster("local")
in Spark application code, the application exit abnormally after about 2 minutes. In driver's
log there is an error whose content is "SparkContext did not initialize after waiting for
100000 ms. Please check earlier log output for errors. Failing the application".
> When SparkContext is launched, it should be checked whether the option --master of spark-submit
script and SparkConf.setMaster in Spark application code are different. If they are different,
there should be a clear hint in the driver's log for the developer to troubleshoot.
> I found the same question with me in stackoverflow:
> http://stackoverflow.com/questions/30670933/submit-spark-job-on-yarn-cluster



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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


Mime
View raw message