[
https://issues.apache.org/jira/browse/MAPREDUCE-5265?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jason Lowe updated MAPREDUCE-5265:
----------------------------------
Status: Open (was: Patch Available)
The patch has gone stale and needs to be refreshed. Other comments:
General:
* protocol files are in package org.apache.hadoop.yarn but
org.apache.hadoop.mapreduce is more appropriate. The history server is
specific to MapReduce and not a generic YARN history server. In the future
YARN may provide its own, separate history server
* Formatting for 80 columns
log4j.properties:
* Using {{yarn.hs}} instead of {{mapreduce.hs}}
I'm torn on whether the history server should be using YARN RPC or just go
straight to Hadoop Common RPC like what was done for the equivalent
functionality in the namenode (see NameNodeRpcServer,
RefreshUserMappingsProtocolServerSideTranslatorPB, etc.). As the MRAppMaster
is already using YARN RPC for its stuff I could see it going either way. Do
others have an opinion on this?
> Need an admin interface on history server with the ability to refresh super
> user groups,refresh user to group mappings,refresh admin acls,get groups
> given a username.
> ----------------------------------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: MAPREDUCE-5265
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-5265
> Project: Hadoop Map/Reduce
> Issue Type: New Feature
> Components: jobhistoryserver
> Affects Versions: 2.1.0-beta
> Reporter: Jason Lowe
> Assignee: Ashwin Shankar
> Attachments: JHS_REFRESH-2.txt, JHS_REFRESH-4.txt, JHS_REFRESH-6.txt
>
>
> The history server needs an admin interface with the ability to
> 1. refresh the super user groups configurations,
> 2. refresh user to group mappings,
> 3. refresh its admin acls,
> 4. get groups given a username
> without requiring a restart of the history server. This is analogous to the
> -refreshSuperUserGroupsConfiguration capabilities provided by hdfs dfsadmin
> and yarn rmadmin.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
