hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sergey Shelukhin (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HIVE-15541) Hive OOM when ATSHook enabled and ATS goes down
Date Fri, 06 Jan 2017 00:22:58 GMT

    [ https://issues.apache.org/jira/browse/HIVE-15541?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15803004#comment-15803004
] 

Sergey Shelukhin commented on HIVE-15541:
-----------------------------------------

Hmm. Why not change the queue to reference requests created at enqueue time, instead of capturing
huge amount of context in the runnable (unless I'm reading this wrong)?
In addition to queue limit.
That way queue limit could probably be set to a larger value.


> Hive OOM when ATSHook enabled and ATS goes down
> -----------------------------------------------
>
>                 Key: HIVE-15541
>                 URL: https://issues.apache.org/jira/browse/HIVE-15541
>             Project: Hive
>          Issue Type: Bug
>          Components: Hooks
>            Reporter: Jason Dere
>            Assignee: Jason Dere
>         Attachments: HIVE-15541.1.patch, HIVE-15541.2.patch
>
>
> The ATS API used by the Hive ATSHook is a blocking call, if ATS goes down this can block
the ATSHook executor, while the hook continues to submit work to the executor with each query.
> Over time the buildup of queued items can cause OOM.



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

Mime
View raw message