You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@hive.apache.org by "Thejas M Nair (JIRA)" <ji...@apache.org> on 2015/08/04 21:08:04 UTC

[jira] [Updated] (HIVE-11407) JDBC DatabaseMetaData.getTables with large no of tables call leads to HS2 OOM

     [ https://issues.apache.org/jira/browse/HIVE-11407?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Thejas M Nair updated HIVE-11407:
---------------------------------
    Description: 
With around 7000 tables having around 1500 columns each, and 512MB of HS2 memory, I am able to reproduce this OOM .
Most of the memory is consumed by the datanucleus objects. 

Reducing the number of tables for which metadata is fetched at a time via metastore get_multi_table limits the memory footprint.

  was:
With around 7000 tables having around 1500 columns each, and 512MB of HS2 memory, I am able to reproduce this OOM .
Most of the memory is consumed by the datanucleus objects. 


> JDBC DatabaseMetaData.getTables with large no of tables call leads to HS2 OOM 
> ------------------------------------------------------------------------------
>
>                 Key: HIVE-11407
>                 URL: https://issues.apache.org/jira/browse/HIVE-11407
>             Project: Hive
>          Issue Type: Bug
>          Components: HiveServer2
>            Reporter: Thejas M Nair
>            Assignee: Sushanth Sowmyan
>         Attachments: HIVE-11407-branch-1.0.patch, HIVE-11407.1.patch
>
>
> With around 7000 tables having around 1500 columns each, and 512MB of HS2 memory, I am able to reproduce this OOM .
> Most of the memory is consumed by the datanucleus objects. 
> Reducing the number of tables for which metadata is fetched at a time via metastore get_multi_table limits the memory footprint.



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