hadoop-mapreduce-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mahadev konar (Commented) (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (MAPREDUCE-3380) Token infrastructure for running clients which are not kerberos authenticated
Date Fri, 11 Nov 2011 22:12:52 GMT

    [ https://issues.apache.org/jira/browse/MAPREDUCE-3380?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13148776#comment-13148776
] 

Mahadev konar commented on MAPREDUCE-3380:
------------------------------------------

@Vinod,

bq. To communicate to the JobHistory server, again we need some delegation token. I guess
we can use the mapreduce JobToken itself but AM somehow needs to let the JHS know about it
- through JobHistory file, or a direct communication protocol.

Do we really need the JHS token? As far as I understand the getDelgationtoken is used by a
mapred task to submit another job to the cluster. Does it check for job status from JHS as
well? In that case we just need to return the application token generated in AMLauncher (probably
need to create that early on) no? Also, do we need the application token to be based on app
attemptid? Can we just use the application id? 


                
> Token infrastructure for running clients which are not kerberos authenticated
> -----------------------------------------------------------------------------
>
>                 Key: MAPREDUCE-3380
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-3380
>             Project: Hadoop Map/Reduce
>          Issue Type: Sub-task
>          Components: mr-am
>    Affects Versions: 0.23.0
>            Reporter: Alejandro Abdelnur
>            Priority: Blocker
>             Fix For: 0.23.1
>
>         Attachments: MAPREDUCE-3380.branch-0.23.patch, MAPREDUCE-3380.branch-0.23.patch
>
>
> The JobClient.getDelegationToken() method is returning NULL, this makes Oozie fail when
trying to get the delegation token to use it for starting a job.
> What is seems to be happing is that Jobclient.getDelegationToken() calls Cluster.getDelegationToken()
that calls YarnRunner.getDelegationToken() that calls ResourceMgrDelegate.getDelegationToken().
And the last one is not implemented. (Thanks Ahmed for tracing this in MR2 code)

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message