[ 
https://issues.apache.org/jira/browse/YARN-6647?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16260882#comment-16260882
 ] 

Jason Lowe commented on YARN-6647:
----------------------------------

Thanks for the patch!

I agree with [~rohithsharma]'s comments.  If the RMDelegationTokenSecretManager 
receives an interrupted exception, either on the expired token remover thread 
or on another thread, it should not cause the JVM to exit if it is being 
shutdown.  The other methods should be updated to handle interrupt exceptions 
similarly even if they won't involve the expired token remover thread.  The 
state store itself could throw an interrupt exception, and the secret manager 
should not tear down the JVM if it knows this interrupt exception is likely a 
side effect of being shut down.

For the AsyncDispatcher change, the new wording is more confusing and doesn't 
add much value.  If the async dispatcher is seeing an InterruptedException then 
the thread is being interrupted.  The original wording conveys that 
effectively.  The new wording suggests a possible reason where the interrupt 
was recognized, but the stack trace will show where the interrupt exception 
originated.  There are lots of methods besides queue put that will check for a 
thread interrupted status and throw the exception, so I do not see a reason to 
call out a specific one here especially when the exception stack trace will 
show from whence it was thrown.


> RM can crash during transitionToStandby due to InterruptedException
> -------------------------------------------------------------------
>
>                 Key: YARN-6647
>                 URL: https://issues.apache.org/jira/browse/YARN-6647
>             Project: Hadoop YARN
>          Issue Type: Bug
>          Components: resourcemanager
>    Affects Versions: 3.0.0-alpha4
>            Reporter: Jason Lowe
>            Assignee: Bibin A Chundatt
>            Priority: Critical
>         Attachments: YARN-6647.001.patch, YARN-6647.002.patch
>
>
> Noticed some tests were failing due to the JVM shutting down early.  I was 
> able to reproduce this occasionally with TestKillApplicationWithRMHA.  
> Stacktrace to follow.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

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

Reply via email to