You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "anishek (JIRA)" <ji...@apache.org> on 2017/06/14 09:31:00 UTC
[jira] [Updated] (HIVE-16897) repl load does not lead to excessive
memory consumption for multiple functions from same binary jar
[ https://issues.apache.org/jira/browse/HIVE-16897?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
anishek updated HIVE-16897:
---------------------------
Issue Type: Sub-task (was: Improvement)
Parent: HIVE-16865
> repl load does not lead to excessive memory consumption for multiple functions from same binary jar
> ----------------------------------------------------------------------------------------------------
>
> Key: HIVE-16897
> URL: https://issues.apache.org/jira/browse/HIVE-16897
> Project: Hive
> Issue Type: Sub-task
> Components: HiveServer2
> Affects Versions: 3.0.0
> Reporter: anishek
> Assignee: anishek
> Fix For: 3.0.0
>
>
> as part of function replication we currently keep a separate copy of the binary jar associated with the function ( this should be same on the primary warehouse also since each hdfs jar location given during creation of function will download the resource in a separate resource location thus leading to the same jar being included in class path multiple times)
> this will lead to excessive space used to keep all jars in classpath, solve this by identifying the common binary jar ( using checksum from primary on replica) and not creating multiple copies thus preventing excessive memory usage.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)