[ 
https://issues.apache.org/jira/browse/SYSTEMML-930?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15502250#comment-15502250
 ] 

Arvind Surve commented on SYSTEMML-930:
---------------------------------------

Setting MaxResultSize after SparkContext started does not get reflected. This 
to be set before starting SparkContext. So if user has not set earlier, warn 
user for the same.

> Warn user to Set MaxResultSize to unlimited in MLContext based call
> -------------------------------------------------------------------
>
>                 Key: SYSTEMML-930
>                 URL: https://issues.apache.org/jira/browse/SYSTEMML-930
>             Project: SystemML
>          Issue Type: Improvement
>            Reporter: Arvind Surve
>            Assignee: Arvind Surve
>
> Currently MaxResultSize has not set in the case MlContext call being made. In 
> case user does not specify it through spark.driver.maxResultSize, it gets set 
> to default size (1g). In order to get unlimited data to driver we have set 
> this size to unlimited in other calls like DML script calls through 
> spark-submit. We need to do similar for MlContext based call.



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

Reply via email to