[ 
https://issues.apache.org/jira/browse/YARN-4479?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15066282#comment-15066282
 ] 

Rohith Sharma K S commented on YARN-4479:
-----------------------------------------

Scenario which cause issue is
# Submitted the app-1 and app-2 with priority 5. Both applications are 
activated and RUNNING state.
# Submit app-3 with priority 6. This application is in pending state because of 
AMLimit.
# RM restarted, app-1 application is activated(it is behavior that for 1st 
application AMLimit is not considered) and app-2 and app-3 are in 
pendingOrderingPolicy
# AM re-registered for app-1 and app-2. Its state is now RUNNING. But app-2 and 
app-3 are still in pendingapplications.
# NodeManager re-registered with RM. As a result 1 application supposed to be 
get activated. Here, always app-3 get activated since app-3 priority is higher, 
but app-2 should get activated first since it is running before RMrestart.

> Retrospect app-priority in pendingOrderingPolicy during recovering 
> applications
> -------------------------------------------------------------------------------
>
>                 Key: YARN-4479
>                 URL: https://issues.apache.org/jira/browse/YARN-4479
>             Project: Hadoop YARN
>          Issue Type: Sub-task
>          Components: api, resourcemanager
>            Reporter: Rohith Sharma K S
>            Assignee: Rohith Sharma K S
>         Attachments: 0001-YARN-4479.patch
>
>
> Currently, same ordering policy is used for pending applications and active 
> applications. When priority is configured for an applications, during 
> recovery high priority application get activated first. It is possible that 
> low priority job was submitted and running state. 
> This causes low priority job in starvation after recovery



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

Reply via email to