You are viewing a plain text version of this content. The canonical link for it is here.
Posted to mapreduce-issues@hadoop.apache.org by "Siddharth Seth (JIRA)" <ji...@apache.org> on 2012/06/29 00:15:44 UTC

[jira] [Commented] (MAPREDUCE-4014) Make yarn-common services robust

    [ https://issues.apache.org/jira/browse/MAPREDUCE-4014?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13403517#comment-13403517 ] 

Siddharth Seth commented on MAPREDUCE-4014:
-------------------------------------------

Steve, apologies for the delayed feedback. I was wondering which parts of MAPREDUCE-3939 are planned for this patch. 
"Static methods to choreograph of lifecycle operations" seems to be covered.
Will the others, specifically "AbstractService doesn't prevent duplicate state change requests", "AbstractService state change doesn't defend against race conditions", "state model prevents stopped state being entered if you could not successfully start the service" be separate patches? Definitely looks like they should go in.

bq. state model prevents stopped state being entered if you could not successfully start the service.
I don't believe resources which require an explicit release are meant to be obtained in the init() stage - but that may not always be the case. I'd agree with allowing stop() from any state, as well as a CompositeService attempting to stop all child services when told to stop() (instead of just the ones which have started).
Currently, a failed start() on a composite service will stop() services which had started and move them to STOPPED state, attempt to stop() the failed service but leave it in INITED state, and leave remaining services in INITED state - which doesn't seem correct.

Feedback on this patch
 - {{interruptAndJoinThread(Thread target)}} - could have a joinTimeout parameter as well.
 - stopIPCServer, stopWebApp, interrupt* - should these be in AbstractService ? or a separate helper class.
 - toString() in AbstractService - the text is missing a closing quote.
                
> Make yarn-common services robust
> --------------------------------
>
>                 Key: MAPREDUCE-4014
>                 URL: https://issues.apache.org/jira/browse/MAPREDUCE-4014
>             Project: Hadoop Map/Reduce
>          Issue Type: Sub-task
>          Components: mrv2
>    Affects Versions: 0.23.1, 0.24.0
>            Reporter: Steve Loughran
>            Assignee: Steve Loughran
>              Labels: yarn
>         Attachments: MAPREDUCE-4014.patch
>
>
> Review the yarn common services ({{CompositeService}}, {{AbstractLivelinessMonitor}} and make their service startup _and especially shutdown_ more robust against out-of-lifecycle invocation and partially complete initialization.
> Write tests for these where possible. 

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira