[jira] [Comment Edited] (HIVE-21642) Hive server leaks memory on data insertion

["Gopal V (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/HIVE-21642?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16823524#comment-16823524 ] 

Gopal V edited comment on HIVE-21642 at 4/22/19 10:38 PM:
----------------------------------------------------------

This does not leak memory on Hive master anymore (~5 min expiry), I recommend opening a ticket with EMR for 2.3.x branch fixes.


was (Author: gopalv):
This does not leak memory on Hive master anymore, I recommend opening a ticket with EMR for 2.3.x branch fixes.

> Hive server leaks memory on data insertion
> ------------------------------------------
>
>                 Key: HIVE-21642
>                 URL: https://issues.apache.org/jira/browse/HIVE-21642
>             Project: Hive
>          Issue Type: Bug
>          Components: HiveServer2
>    Affects Versions: 2.3.4
>         Environment: * Amazon Hadoop Distribution emr-5.20.0
>  * Master mode with 4 CPU and 16 GB RAM each
>  * Table files stored in S3 cloud storage
>            Reporter: Alexander Knopov
>            Priority: Major
>
> We are continuously loading data into Hive table stored in ORC format by appending data in batches. We repeatedly have seen that over a span of few days Hive server experience {{OutOfMemoryError}} exceptions that we believe are caused by memory leaks.
> Comparing heap dumps shows that most suspicious classes that show persistent growth are and not recycled with GC are
>  * {{org.apache.hadoop.hive.ql.io.orc.OrcStruct$Field}}
>  * {{org.apache.hadoop.hive.serde2.objectinspector.UnionStructObjectInspector$MyField}}
>  * {{String}}
> Sample program used for stress test and heap dumps from 700 to 2500 GB can be uploaded on request. They are too big for Jira backing store



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)