hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Hadoop QA (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-2134) FairSchduler AllocationConfiguration doesn't treat hierarchical queues settings
Date Mon, 09 Jun 2014 10:13:01 GMT

    [ https://issues.apache.org/jira/browse/YARN-2134?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14021828#comment-14021828

Hadoop QA commented on YARN-2134:

{color:red}-1 overall{color}.  Here are the results of testing the latest attachment 
  against trunk revision .

    {color:red}-1 patch{color}.  The patch command could not apply the patch.

Console output: https://builds.apache.org/job/PreCommit-YARN-Build/3939//console

This message is automatically generated.

> FairSchduler AllocationConfiguration doesn't treat hierarchical queues settings
> -------------------------------------------------------------------------------
>                 Key: YARN-2134
>                 URL: https://issues.apache.org/jira/browse/YARN-2134
>             Project: Hadoop YARN
>          Issue Type: Bug
>          Components: fairscheduler
>    Affects Versions: 2.5.0
>            Reporter: Andrey Stepachev
>         Attachments: YARN-2134.diff, broken-hier-queues.diff
> AllocationConfiguration doesn't fully qualify queue names with respect to 'root.' prefix.
Such behaviour leads to situation where settings for nested queue doesn't take effect. 
> Consider following config (working example in broken-hier-queues.diff attached) sub queue
settings were not applied by method queueConf.getMaxResources.
> {code:xml}
> <queue name="queueA">
> <maxResources>2048mb,10vcores</maxResources>
> <queue name="queueB">
> <maxResources>512mb,1vcores</maxResources>
> </queue>
> </queue>
> {code}
> I made patch, which unifies QueueName operations (remove those indexOf copy pastes) and
apply some of settings in hierarchical manner  (say preemption timeout can be used from parent
queue instead of global default).
> Also it is interesting to introduce type safe object QueueName and replace String in
interfaces, and I have made an attempt (https://github.com/octo47/hadoop-common/tree/feature/octo/fs-queue-name),
but looks like it lead to some inconsitancy and I think that it is not necessary right now).

This message was sent by Atlassian JIRA

View raw message