[
https://issues.apache.org/jira/browse/AMBARI-19116?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15843419#comment-15843419
]
Swapan Shridhar commented on AMBARI-19116:
------------------------------------------
commit:
branch-2.5:
{code}
commit 8552345deee7e6d366571a4c926a3093f7a7492a
Author: Sumit Mohanty <[email protected]>
Date: Sun Dec 11 11:15:46 2016 -0800
AMBARI-18901. LLAP integration enhancements (Swapan Sridhar via smohanty)
{code}
trunk:
{code}
commit 8552345deee7e6d366571a4c926a3093f7a7492a
Author: Sumit Mohanty <[email protected]>
Date: Sun Dec 11 11:15:46 2016 -0800
AMBARI-18901. LLAP integration enhancements (Swapan Sridhar via smohanty)
{code}
> Hive2/HiveServer2 (HSI) heap size calculation to be based on LLAP concurrency.
> ------------------------------------------------------------------------------
>
> Key: AMBARI-19116
> URL: https://issues.apache.org/jira/browse/AMBARI-19116
> Project: Ambari
> Issue Type: Bug
> Components: ambari-server
> Affects Versions: 2.4.1
> Reporter: Swapan Shridhar
> Assignee: Swapan Shridhar
> Priority: Critical
> Fix For: 2.5.0
>
> Attachments: AMBARI-19116.patch
>
>
> - Currently, hive2/HiveServer2 (HSI) heap size is ket same as that of
> hive/HiveServer2.
> Required:
> - Hive2/HiveServer2 (HSI) heap size calculations to be based on LLAP
> concurrency.
> {code}
> = min(max(2GB, 400*LLAP concurrency (Tez sessions)), 3/8 * <HS2 host total
> RAM>)
> {code}
> *Side fix:* Also, updated "hive.llap.daemon.queue.name" and
> "hive.server2.tez.default.queues" config values to one of the existing leaf
> queues, in case of HSI being turned on and config values have queue names
> which no more exist. The scenario can happen when, say :
> - HSI was ON and the above 2 configs has queue name 'XYZ'.
> - HSI was then turned off.
> - Capacity Scheduler queues were changed and 'XYZ' queue was deleted.
> - The above configs without this fix will otherwise show stale/non-existing
> queue name.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
