[
https://issues.apache.org/jira/browse/YARN-1867?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vinod Kumar Vavilapalli updated YARN-1867:
------------------------------------------
Attachment: YARN-1867-20140325.txt
The problem is that the web-services cache the acls-managers from the
'previous' RM. The acls-manager are recreated when a transition happens.
Here's a patch to fix the issue
- Changed web-services to not cache the application and queue acls-managers. I
checked other instances in the web-app. These seem like the only two cached
objects.
- The code in the main ResourceManager has become unmaintenable after the
introduction of the active-services. I had to resist cleaning up, quite a few
things are broken in more ways than one. For now, moved a couple of things from
the top level to be nested inside active-services. Will file a ticket for more
cleanup.
- Fixed few existing formatting issues
- The test case fails without the code change with the same exception printed
above and passes with.
> NPE while fetching apps via the REST API
> ----------------------------------------
>
> Key: YARN-1867
> URL: https://issues.apache.org/jira/browse/YARN-1867
> Project: Hadoop YARN
> Issue Type: Bug
> Components: resourcemanager
> Affects Versions: 2.4.0
> Reporter: Karthik Kambatla
> Assignee: Karthik Kambatla
> Priority: Blocker
> Labels: rest_api
> Attachments: YARN-1867-20140325.txt
>
>
> We ran into the following NPE when fetching applications using the REST API:
> {noformat}
> INTERNAL_SERVER_ERROR
> java.lang.NullPointerException
> at
> org.apache.hadoop.yarn.server.security.ApplicationACLsManager.checkAccess(ApplicationACLsManager.java:104)
> at
> org.apache.hadoop.yarn.server.resourcemanager.webapp.RMWebServices.hasAccess(RMWebServices.java:123)
> at
> org.apache.hadoop.yarn.server.resourcemanager.webapp.RMWebServices.getApps(RMWebServices.java:418)
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.2#6252)
