You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-issues@hadoop.apache.org by "Hitesh Sharma (JIRA)" <ji...@apache.org> on 2016/05/26 16:48:14 UTC

[jira] [Commented] (YARN-5162) Exceptions thrown during AM registerAM call when Distributed Scheduling is Enabled

    [ https://issues.apache.org/jira/browse/YARN-5162?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15302394#comment-15302394 ] 

Hitesh Sharma commented on YARN-5162:
-------------------------------------

On further debugging, there seems to be three issues:

# It is seen that this is caused due to the {{SchedulerSecurityInfo}} class disallowing all protocols except the ApplicationMasterProtocol.
# Once that was fixed, it was noticed that the AM always dies with a {{NullPointerException}}, since the {{DistSchedRegisterResponse}} did not contain a min and incr allocation capability
# Finally, it looks like the {{finishApplicationMaster}} call does not work. This seems to be because the method is not declared in the proto file.

> Exceptions thrown during AM registerAM call when Distributed Scheduling is Enabled
> ----------------------------------------------------------------------------------
>
>                 Key: YARN-5162
>                 URL: https://issues.apache.org/jira/browse/YARN-5162
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>            Reporter: Arun Suresh
>            Assignee: Hitesh Sharma
>
> The following Exception is seen and the AM fails to register with RM:
> {noformat}
> 16/05/24 17:09:26 INFO ipc.Server: Auth successful for appattempt_1464108777756_0001_000001 (auth:SIMPLE)
> 16/05/24 17:09:26 INFO amrmproxy.AMRMProxyService: Registering application master. Host: Port:0 Tracking Url:
> 16/05/24 17:09:26 INFO amrmproxy.DefaultRequestInterceptor: Forwarding registration request to the real YARN RM
> 16/05/24 17:09:26 DEBUG nodemanager.NodeStatusUpdaterImpl: Node's health-status : true,
> 16/05/24 17:09:26 DEBUG nodemanager.NodeStatusUpdaterImpl: Sending out 1 container statuses: [ContainerStatus: [ContainerId: container_1464108777756_0001_01_000
> 001, ExecutionType: GUARANTEED, State: RUNNING, Capability: <memory:1024, vCores:1>, Diagnostics: , ExitStatus: -1000, ]]
> 16/05/24 17:09:26 WARN ipc.Client: Exception encountered while connecting to the server : org.apache.hadoop.security.AccessControlException: Client cannot authe
> nticate via:[TOKEN]
> {noformat}



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org