You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-dev@hadoop.apache.org by "Daniel Templeton (JIRA)" <ji...@apache.org> on 2017/03/07 19:38:38 UTC

[jira] [Created] (YARN-6299) FairSharePolicy is off when demand is less than min share

Daniel Templeton created YARN-6299:
--------------------------------------

             Summary: FairSharePolicy is off when demand is less than min share
                 Key: YARN-6299
                 URL: https://issues.apache.org/jira/browse/YARN-6299
             Project: Hadoop YARN
          Issue Type: Bug
          Components: fairscheduler
    Affects Versions: 3.0.0-alpha2
            Reporter: Daniel Templeton


{code}
      Resource resourceUsage1 = s1.getResourceUsage();
      Resource resourceUsage2 = s2.getResourceUsage();
      Resource minShare1 = Resources.min(RESOURCE_CALCULATOR, null,
          s1.getMinShare(), s1.getDemand());
      Resource minShare2 = Resources.min(RESOURCE_CALCULATOR, null,
          s2.getMinShare(), s2.getDemand());
      boolean s1Needy = Resources.lessThan(RESOURCE_CALCULATOR, null,
          resourceUsage1, minShare1);
      boolean s2Needy = Resources.lessThan(RESOURCE_CALCULATOR, null,
          resourceUsage2, minShare2);
      minShareRatio1 = (double) resourceUsage1.getMemorySize()
          / Resources.max(RESOURCE_CALCULATOR, null, minShare1, ONE).getMemorySize();
      minShareRatio2 = (double) resourceUsage2.getMemorySize()
          / Resources.max(RESOURCE_CALCULATOR, null, minShare2, ONE).getMemorySize();
{code}

If demand is less than min share, then an app will be flagged as needy if it has demand that is higher than its usage, which happens any time the app has been assigned resources that it hasn't started using yet.  That sounds wrong to me.  [~kasha], [~yufeigu]?



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-dev-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-dev-help@hadoop.apache.org