hadoop-yarn-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Carlo Curino (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (YARN-1471) The SLS simulator is not running the preemption policy for CapacityScheduler
Date Fri, 06 Dec 2013 23:05:35 GMT

    [ https://issues.apache.org/jira/browse/YARN-1471?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13841849#comment-13841849

Carlo Curino commented on YARN-1471:

Alejandro, I tried the second version of what you proposed. It does not crash and seem to
sort of work, but talking with Chris it has some very bad properties. If the inner CapacityScheduler
(now or in the future) invokes from within one of its methods any of the methods you guys
want to trap, this implementation would completely miss it. 

This is also true for the ResourceWrapper more generally I think, so probably the trick of
subclassing instead of wrapping would be safer to propagate to the FairScheduler side as well.
The maintenance cost here is that when a new scheduler is added, you have to create another
one of this sub classes, but this seems rare enough of an event to be ok.

> The SLS simulator is not running the preemption policy for CapacityScheduler
> ----------------------------------------------------------------------------
>                 Key: YARN-1471
>                 URL: https://issues.apache.org/jira/browse/YARN-1471
>             Project: Hadoop YARN
>          Issue Type: Bug
>            Reporter: Carlo Curino
>            Assignee: Carlo Curino
>            Priority: Minor
>         Attachments: SLSCapacityScheduler.java, YARN-1471.patch, YARN-1471.patch.2
> The simulator does not run the ProportionalCapacityPreemptionPolicy monitor.  This is
because the policy needs to interact with a CapacityScheduler, and the wrapping done by the
simulator breaks this. 

This message was sent by Atlassian JIRA

View raw message