hive-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Xuefu Zhang (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HIVE-8868) SparkSession and SparkClient mapping[Spark Branch]
Date Fri, 14 Nov 2014 15:21:34 GMT

    [ https://issues.apache.org/jira/browse/HIVE-8868?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14212371#comment-14212371
] 

Xuefu Zhang commented on HIVE-8868:
-----------------------------------

Seems similar to HIVE-7593. Nevertheless, spark client should be user-session specific. However,
in case of local mode, user sessions "should" share the same SparkContext instance, due to
the concurrency problem in SparkContext.

> SparkSession and SparkClient mapping[Spark Branch]
> --------------------------------------------------
>
>                 Key: HIVE-8868
>                 URL: https://issues.apache.org/jira/browse/HIVE-8868
>             Project: Hive
>          Issue Type: Sub-task
>          Components: Spark
>            Reporter: Chengxiang Li
>              Labels: Spark-M3
>
> It should be a seperate spark context for each user session, currently we share a singleton
local spark context in all user sessions with local spark, and create remote spark context
for each spark job with spark cluster.
> To binding one spark context to each user session, we may construct spark client on session
open, one thing to notify is that, is SparkSession::conf is consist with Context::getConf?




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

Mime
View raw message