[
https://issues.apache.org/jira/browse/YARN-621?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13647687#comment-13647687
]
Allen Wittenauer commented on YARN-621:
---------------------------------------
Some more details:
* With the exception of fixing the broken start-dfs.sh, this is a pure Apache
2.0.4 deploy on RHEL 6.3.
* We configure logical nics and bind services to them. In order to work around
HADOOP-9520, we have hard coded all the service names in the configuration.
* After the first job, authentication works and the system works as expected.
If that jobs rolls off the page, the replay errors return.
* Same realm or cross-realm does not appear to make a difference.
* The only stack trace I'm able to find is the one generated by the filter for
the replay error itself.
* Hit this in both Firefox and Safari (which also triggers HADOOP-9521).
> RM triggers web auth failure before first job
> ---------------------------------------------
>
> Key: YARN-621
> URL: https://issues.apache.org/jira/browse/YARN-621
> Project: Hadoop YARN
> Issue Type: Bug
> Components: resourcemanager
> Affects Versions: 2.0.4-alpha
> Reporter: Allen Wittenauer
> Assignee: Vinod Kumar Vavilapalli
> Priority: Critical
>
> On a secure YARN setup, before the first job is executed, going to the web
> interface of the resource manager triggers authentication errors.
--
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
