[
https://issues.apache.org/jira/browse/HIVE-15159?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15660930#comment-15660930
]
Lefty Leverenz commented on HIVE-15159:
---------------------------------------
Doc note: This changes the default value of
*hive.llap.daemon.memory.per.instance.mb* which was introduced by HIVE-11358
and isn't documented in the wiki yet. It belongs in one of the LLAP
subsections in Configuration Properties.
* [Configuration Properties -- LLAP |
https://cwiki.apache.org/confluence/display/Hive/Configuration+Properties#ConfigurationProperties-LLAP]
Added a TODOC2.2 label.
> LLAP ContainerRunner should not reduce the available heap while distributing
> to individual executors
> ----------------------------------------------------------------------------------------------------
>
> Key: HIVE-15159
> URL: https://issues.apache.org/jira/browse/HIVE-15159
> Project: Hive
> Issue Type: Task
> Components: llap
> Reporter: Siddharth Seth
> Assignee: Siddharth Seth
> Labels: TODOC2.2
> Fix For: 2.2.0
>
> Attachments: HIVE-15159.01.patch
>
>
> hive.llap.daemon.memory.per.instance.mb is set to the heapsize of the process
> when llap service driver is used.
> ContainerRunner currently cuts this down to 80% of the available. This
> reduction is not necessary if the heapsize is used, instead of a container
> size.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
