[ 
https://issues.apache.org/jira/browse/YARN-3271?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

nijel updated YARN-3271:
------------------------
    Attachment: YARN-3271.4.patch

Thanks [~kasha] for the comments
 bq.tearDown need not explicitly stop the scheduler. Stopping the RM should 
take care of the scheduler as well.
Done
bq.testNotUserAsDefaultQueue and testDontAllowUndeclaredPools need not stop the 
RM and re-instantiate it. We could just call scheduler.reinitialize
I tried this. As per my analysis the reinitialize call will not consider the 
conf object.  
{code}
FairScheduler.java

  @Override
  public void reinitialize(Configuration conf, RMContext rmContext)
      throws IOException {
    try {
      allocsLoader.reloadAllocations();
    } catch (Exception e) {
      LOG.error("Failed to reload allocations file", e);
    }
  }
Here conf is not used.
{code}
bq.testMoveRunnableApp - Remove commented out scheduler.init and start
Done

> FairScheduler: Move tests related to max-runnable-apps from TestFairScheduler 
> to TestAppRunnability
> ---------------------------------------------------------------------------------------------------
>
>                 Key: YARN-3271
>                 URL: https://issues.apache.org/jira/browse/YARN-3271
>             Project: Hadoop YARN
>          Issue Type: Improvement
>            Reporter: Karthik Kambatla
>            Assignee: nijel
>              Labels: BB2015-05-TBR
>         Attachments: YARN-3271.1.patch, YARN-3271.2.patch, YARN-3271.3.patch, 
> YARN-3271.4.patch
>
>




--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to