[
https://issues.apache.org/jira/browse/HIVE-19763?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16523458#comment-16523458
]
Peter Vary commented on HIVE-19763:
-----------------------------------
[~sershe]: The above mentioned WM is the one under this umbrella jira:
HIVE-17481? Am I right, when I think that WM helps when we have an already
compiled query and we want them to gracefully handle cluster resources when
executing the given query (and other queries in parallel)?
I think [~lmarti...@cloudera.com]'s problem is more related to HS2 memory
management, where compiling and optimizing the query causes OOM on HS2.
> Prevent execution of very large queries
> ---------------------------------------
>
> Key: HIVE-19763
> URL: https://issues.apache.org/jira/browse/HIVE-19763
> Project: Hive
> Issue Type: New Feature
> Components: HiveServer2
> Affects Versions: 2.3.2
> Reporter: Luis E Martinez-Poblete
> Priority: Minor
>
> Synopsis:
> =========
> Prevent execution of very large queries.
>
> Feature Request:
> ================
> Please enhance Hive with a parameter to restrict the execution of very large
> queries.
> Use case: User is trying to create a view with a size of 8 MB. Creation of
> this view was possible after increasing heap memory in several components
> (HMS, HS2, Zookeeper). However, this view caused major issues when it was
> used in a CTE query which resulted in GC pauses and eventually OOM of the HS2
> process.
>
> Although, it is possible to create the view, it may cause other issues when
> used in queries. From the Hadoop administrator point of view, it would be
> good to restrict this type of queries.
>
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
