[jira] [Commented] (SAMZA-382) SamzaContainer main thread might die without proper cleanup

["Chris Riccomini (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SAMZA-382?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14097841#comment-14097841 ] 

Chris Riccomini commented on SAMZA-382:
---------------------------------------

bq. The result is that, the AMF thread can still be alive - thus locking up resources in YARN.

One minor note: the thread is the JMX server (non-daemon) thread, not the AMF one (which is an internal metrics thing at LI :)).

> SamzaContainer main thread might die without proper cleanup
> -----------------------------------------------------------
>
>                 Key: SAMZA-382
>                 URL: https://issues.apache.org/jira/browse/SAMZA-382
>             Project: Samza
>          Issue Type: Bug
>          Components: container
>    Affects Versions: 0.8.0
>            Reporter: Chinmay Soman
>             Fix For: 0.8.0
>
>
> In SamzaContainer, 
> There are some method invocations within 'main' that throw an exception:
> For eg: getTaskNameToSystemStreamPartition (this might happen if there is no SSP for a given task). 
> However, this is outside the try block due to which, it doesn't cleanup properly before dying. The result is that, the AMF thread can still be alive - thus locking up resources in YARN.
> Proposed resolution: Include all such statements within the try block



--
This message was sent by Atlassian JIRA
(v6.2#6252)