[
https://issues.apache.org/jira/browse/HIVE-13240?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15186376#comment-15186376
]
Gopal V commented on HIVE-13240:
--------------------------------
[~wyukawa]: take a look at this patch for your Tez OOM issue.
My primary suspect is that you need HIVE-10529 to fix that issue, but to
confirm this issue or the other - you need to confirm that this issue goes away
when you do
hive --hiveconf tez.am.container.reuse.enabled=false;
(note, has to be in the cli --hiveconf, doing set in CLI doesn't work)
> GroupByOperator: Drop the hash aggregates when closing operator
> ---------------------------------------------------------------
>
> Key: HIVE-13240
> URL: https://issues.apache.org/jira/browse/HIVE-13240
> Project: Hive
> Issue Type: Bug
> Components: Query Processor
> Affects Versions: 1.3.0, 1.2.1, 2.0.0
> Reporter: Gopal V
> Assignee: Gopal V
> Attachments: HIVE-13240.1.patch
>
>
> GroupByOperator holds onto the Hash aggregates accumulated when the plan is
> cached.
> Drop the hashAggregates in case of error during forwarding to the next
> operator.
> Added for PTF, TopN and all GroupBy cases.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
