[jira] [Commented] (YARN-300) After yarn-271, when yarn.scheduler.fair.max.assign=0, fairscheduler will infinite loop and not schedule any application.

2012-12-31 Thread caolong (JIRA) 

[ 
https://issues.apache.org/jira/browse/YARN-300?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13541588#comment-13541588
 ] 

caolong commented on YARN-300:
--

That's perfectly correct!! default maxAssign=-1 and assignMultiple=true so 
fairScheduler will infinite loop.

 After yarn-271, when yarn.scheduler.fair.max.assign=0, fairscheduler will  
 infinite loop and not schedule any application.
 --

 Key: YARN-300
 URL: https://issues.apache.org/jira/browse/YARN-300
 Project: Hadoop YARN
  Issue Type: Bug
  Components: resourcemanager, scheduler
Reporter: shenhong
  Labels: None
 Fix For: 2.0.3-alpha

 Attachments: YARN-300.patch

   Original Estimate: 10h
  Remaining Estimate: 10h

 After yarn-271, when yarn.scheduler.fair.max.assign=0, when a node was been 
 reserved, fairScheduler will  infinite loop and not schedule any application.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

[jira] [Commented] (YARN-300) After yarn-271, when yarn.scheduler.fair.max.assign=0, fairscheduler will infinite loop and not schedule any application.

2012-12-31 Thread shenhong (JIRA) 

[ 
https://issues.apache.org/jira/browse/YARN-300?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanelfocusedCommentId=13541590#comment-13541590
 ] 

shenhong commented on YARN-300:
---

At FairScheduler#nodeUpdate() the code is:
if ((assignedContainers = maxAssign)  (maxAssign  0)) { break; }
so when the maxAssign = 0, it will infinite loop, and default maxAssign=-1.

 After yarn-271, when yarn.scheduler.fair.max.assign=0, fairscheduler will  
 infinite loop and not schedule any application.
 --

 Key: YARN-300
 URL: https://issues.apache.org/jira/browse/YARN-300
 Project: Hadoop YARN
  Issue Type: Bug
  Components: resourcemanager, scheduler
Reporter: shenhong
  Labels: None
 Fix For: 2.0.3-alpha

 Attachments: YARN-300.patch

   Original Estimate: 10h
  Remaining Estimate: 10h

 After yarn-271, when yarn.scheduler.fair.max.assign=0, when a node was been 
 reserved, fairScheduler will  infinite loop and not schedule any application.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira