hive-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mithun Radhakrishnan (JIRA)" <j...@apache.org>
Subject [jira] [Comment Edited] (HIVE-17754) InputJobInfo in Pig UDFContext is heavyweight, and causes OOMs in Tez AMs
Date Tue, 10 Oct 2017 17:00:10 GMT

    [ https://issues.apache.org/jira/browse/HIVE-17754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16198244#comment-16198244
] 

Mithun Radhakrishnan edited comment on HIVE-17754 at 10/10/17 4:59 PM:
-----------------------------------------------------------------------

This fix depends on HIVE-11548. The attached patch contains both the fix for HIVE-11548 and
the one for HIVE-17754. Submitting for tests...


was (Author: mithun):
This fix depends on HIVE-11548. The attached patch contains both the fix for HIVE-11548 and
HIVE-17754. Submitting for tests...

> InputJobInfo in Pig UDFContext is heavyweight, and causes OOMs in Tez AMs
> -------------------------------------------------------------------------
>
>                 Key: HIVE-17754
>                 URL: https://issues.apache.org/jira/browse/HIVE-17754
>             Project: Hive
>          Issue Type: Bug
>          Components: HCatalog
>    Affects Versions: 2.2.0, 3.0.0
>            Reporter: Mithun Radhakrishnan
>            Assignee: Mithun Radhakrishnan
>         Attachments: HIVE-17754.1.patch
>
>
> HIVE-9845 dealt with reducing the size of HCat split-info, to improve job-launch times
for Pig/HCat jobs.
> For large Pig queries that scan a large number of Hive partitions, it was found that
the Pig {{UDFContext}} stored full-fat HCat {{InputJobInfo}} objects, thus blowing out the
Pig Tez AM. Since this information is already stored in the {{HCatSplit}}, the serialization
of {{InputJobInfo}} can be spared.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message