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-3800) Simplify inmemory state for ReservationAllocation
Date Wed, 08 Jul 2015 23:08:05 GMT

    [ https://issues.apache.org/jira/browse/YARN-3800?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14619550#comment-14619550

Carlo Curino commented on YARN-3800:

Patch generally looks good, and I spoke with [~subru] that explained me why you are making
these changes, and overall makes sense. 

A couple nits and then I am ok to commit this:
1) I think it is nicer to have types in HashMap<> and TreeMap<> initializations.
2) In other places you did this change already, but in TestRLESparseResourceAllocation  you
have a generateAllocation that still produces ReservationRequests and then you immediately
convert to Resource. Probably easier to change generateAllocation

Thanks for the work on this patch.

> Simplify inmemory state for ReservationAllocation
> -------------------------------------------------
>                 Key: YARN-3800
>                 URL: https://issues.apache.org/jira/browse/YARN-3800
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: capacityscheduler, fairscheduler, resourcemanager
>            Reporter: Anubhav Dhoot
>            Assignee: Anubhav Dhoot
>         Attachments: YARN-3800.001.patch, YARN-3800.002.patch, YARN-3800.002.patch, YARN-3800.003.patch,
> Instead of storing the ReservationRequest we store the Resource for allocations, as thats
the only thing we need. Ultimately we convert everything to resources anyway

This message was sent by Atlassian JIRA

View raw message