[
https://issues.apache.org/jira/browse/MAPREDUCE-3007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13105079#comment-13105079
]
Mahadev konar commented on MAPREDUCE-3007:
------------------------------------------
Vinod,
The conf property setting is not needed any longer:
{code}
myConf.setClass(YarnConfiguration.YARN_SECURITY_INFO,
- ClientRMSecurityInfo.class, SecurityInfo.class);
- YarnRPC rpc = YarnRPC.create(myConf);
+ ClientHSSecurityInfo.class, SecurityInfo.class);
{code}
Setting the conf property is redundant since SecurityUtil picks it up via going
through the providers. The only issue with HS is that both the AM and HS
implement the same interface (MRClientProtocol) one of which is token based and
the other kerberos info based (the fact that getKerberos in
MRClientSecurityInfo returns null saves the day).
I tried out the patch on a cluster. Had to remove the sleep in AppMaster to
reproduce the issue. The patch worked for me, though I started getting some
other error on getTaskCompletions(). I think thats more related to the
handshake for history file transfer not getting completed becasue of immediate
AppMaster shutdown.
In all, the patch looks good with minor changes to removing the conf settings
mentioned above.
> JobClient cannot talk to JobHistory server in secure mode
> ---------------------------------------------------------
>
> Key: MAPREDUCE-3007
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-3007
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: jobhistoryserver, mrv2
> Affects Versions: 0.23.0
> Reporter: Vinod Kumar Vavilapalli
> Assignee: Vinod Kumar Vavilapalli
> Fix For: 0.23.0
>
> Attachments: MAPREDUCE-3007-20110914.2.txt,
> MAPREDUCE-3007-20110914.txt
>
>
> In secure mode, Jobclient cannot connect to HistoryServer. Thanks to
> [~karams] for finding this out.
> {code}
> 11/09/14 09:57:51 INFO mapred.ClientServiceDelegate: Application state is
> completed. Redirecting to job history server
> 11/09/14 09:57:51 INFO security.ApplicationTokenSelector: Looking for a token
> with service <history-server>:10020
> 11/09/14 09:57:51 INFO security.ApplicationTokenSelector: Token kind is
> YARN_APPLICATION_TOKEN and the token's service name is <Am-ip>:46257
> 11/09/14 09:57:51 INFO security.UserGroupInformation: Initiating logout for
> <user-principal>
> 11/09/14 09:57:51 INFO security.UserGroupInformation: Initiating re-login for
> <user-principal>
> 11/09/14 09:57:55 WARN security.UserGroupInformation: Not attempting to
> re-login since the last re-login was attempted less than 600 seconds before.
> 11/09/14 09:57:56 WARN security.UserGroupInformation: Not attempting to
> re-login since the last re-login was attempted less than 600 seconds before.
> 11/09/14 09:58:00 WARN security.UserGroupInformation: Not attempting to
> re-login since the last re-login was attempted less than 600 seconds before.
> 11/09/14 09:58:05 WARN security.UserGroupInformation: Not attempting to
> re-login since the last re-login was attempted less than 600 seconds before.
> 11/09/14 09:58:05 WARN ipc.Client: Couldn't setup connection for
> <user-principal> to null
> 11/09/14 09:58:05 INFO mapred.ClientServiceDelegate: Failed to contact
> AM/History for job job_1315993268700_0001 Will retry..
> {code}
> Am surprised no one working with YARN+MR ever ran into this!
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
