[
https://issues.apache.org/jira/browse/YARN-190?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13704036#comment-13704036
]
Omkar Vinit Joshi commented on YARN-190:
----------------------------------------
[~xiajunluan] there were major changes in distributed shell recently and now
there doesn't seem to be any problem for the scenario you have mentioned for
current trunk. If you see this issue again then please reopen the issue. It
would be really helpful if you can upload RM/NM/AM/Container logs as that will
help a lot in debugging. Thanks. closing it as not reproducible.
> Issues when running distributedShell example in hadoop-2.0.1-alpha cluster
> --------------------------------------------------------------------------
>
> Key: YARN-190
> URL: https://issues.apache.org/jira/browse/YARN-190
> Project: Hadoop YARN
> Issue Type: Bug
> Components: applications/distributed-shell
> Affects Versions: 2.0.1-alpha
> Environment: ubuntu 11.04
> Reporter: xiajunluan
> Assignee: Omkar Vinit Joshi
> Priority: Critical
>
> I have successfully run distributed Shell example in single node that
> deployed the hadoop-2.0.1-alpha.
> But when I run distributedShell in cluster environment, sometimes it works
> well, but sometimes it will failed, following is my detail configuration
>
> A: NameNode, ResourceManager
> B: DataNode, NodeManager
> C: DataNode, NodeManager
> I run the distributedShell with command
> “./bin/hadoop jar
> share/hadoop/mapreduce/hadoop-yarn-applications-distributedshell-2.0.1-alpha.jar
> org.apache.hadoop.yarn.applications.distributedshell.Client -jar
> share/hadoop/mapreduce/hadoop-yarn-applications-distributedshell-2.0.1-alpha.jar
> -shell_command whoami -debug”
>
> When application master is launched, it will create the container to run
> shell command “whoami”, so application master will run on node B or C, and
> container will also randomly run on B or C, if application master and shell
> command container run on the same node(for example all on node B), above
> command will run successfully, but if in different node, that is to say, if
> application master launched successfully in node B and it create the
> container that will run on node C, I will receive the error message:
> …….
> 12/10/29 19:18:02 INFO distributedshell.Client: Application did finished
> unsuccessfully. YarnState=FINISHED, DSFinalStatus=FAILED. Breaking monitoring
> loop
> 12/10/29 19:18:02 ERROR distributedshell.Client: Application failed to
> complete successfully”
--
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
