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] [Comment Edited] (HIVE-14433) refactor LLAP plan cache avoidance and fix issue in merge processor
Date Fri, 05 Aug 2016 19:38:20 GMT

    [ https://issues.apache.org/jira/browse/HIVE-14433?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15409954#comment-15409954
] 

Sergey Shelukhin edited comment on HIVE-14433 at 8/5/16 7:37 PM:
-----------------------------------------------------------------

[~prasanth_j] can you take a look? Note this also fixes the usage in Reduce and Merge processors
where close() uses the "cache" field, but that field is never initialized - instead, ctors
use a local variable.


was (Author: sershe):
[~prasanth_j] can you take a look?

> refactor LLAP plan cache avoidance and fix issue in merge processor
> -------------------------------------------------------------------
>
>                 Key: HIVE-14433
>                 URL: https://issues.apache.org/jira/browse/HIVE-14433
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Sergey Shelukhin
>            Assignee: Sergey Shelukhin
>         Attachments: HIVE-14433.patch
>
>
> Map and reduce processors do this:
> {noformat}
>     if (LlapProxy.isDaemon()) {
>       cache = new org.apache.hadoop.hive.ql.exec.mr.ObjectCache(); // do not cache plan
> ...
> {noformat}
> but merge processor just gets the plan. If it runs in LLAP, it can get a cached plan.
Need to move this logic into ObjectCache itself, via a isPlan arg or something. That will
also fix this issue for merge processor



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

Mime
View raw message