You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-issues@hadoop.apache.org by "Thomas Graves (JIRA)" <ji...@apache.org> on 2012/10/22 03:08:11 UTC
[jira] [Created] (YARN-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Thomas Graves created YARN-177:
----------------------------------
Summary: CapacityScheduler - adding a queue while the RM is running has wacky results
Key: YARN-177
URL: https://issues.apache.org/jira/browse/YARN-177
Project: Hadoop YARN
Issue Type: Bug
Components: capacityscheduler
Affects Versions: 0.23.3
Reporter: Thomas Graves
Assignee: Thomas Graves
Priority: Critical
Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
Looking at the RM logs, used memory can go negative but other logs show the number positive:
2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Thomas Graves (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13481529#comment-13481529 ]
Thomas Graves commented on YARN-177:
------------------------------------
I believe the issue here is that the parent of the new queue is not getting set appropriately. When re-initializing it creates an entire new root queue from the configs and then it moves any new queues over to the original root. It doesn't appear to set the parent from the new root back to the original root when it moves it over.
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Thomas Graves
> Priority: Critical
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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] [Updated] (YARN-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Arun C Murthy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Arun C Murthy updated YARN-177:
-------------------------------
Attachment: YARN-177.patch
Good catch Tom, here is a simple fix. I'll test some more and add some unit tests.
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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] [Updated] (YARN-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Arun C Murthy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Arun C Murthy updated YARN-177:
-------------------------------
Attachment: YARN-177.patch
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch, YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13482518#comment-13482518 ]
Hadoop QA commented on YARN-177:
--------------------------------
{color:red}-1 overall{color}. Here are the results of testing the latest attachment
http://issues.apache.org/jira/secure/attachment/12550499/YARN-177.patch
against trunk revision .
{color:green}+1 @author{color}. The patch does not contain any @author tags.
{color:green}+1 tests included{color}. The patch appears to include 1 new or modified test files.
{color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings.
{color:green}+1 javadoc{color}. The javadoc tool did not generate any warning messages.
{color:green}+1 eclipse:eclipse{color}. The patch built with eclipse:eclipse.
{color:red}-1 findbugs{color}. The patch appears to introduce 1 new Findbugs (version 1.3.9) warnings.
{color:green}+1 release audit{color}. The applied patch does not increase the total number of release audit warnings.
{color:green}+1 core tests{color}. The patch passed unit tests in hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager.
{color:green}+1 contrib tests{color}. The patch passed contrib unit tests.
Test results: https://builds.apache.org/job/PreCommit-YARN-Build/118//testReport/
Findbugs warnings: https://builds.apache.org/job/PreCommit-YARN-Build/118//artifact/trunk/patchprocess/newPatchFindbugsWarningshadoop-yarn-server-resourcemanager.html
Console output: https://builds.apache.org/job/PreCommit-YARN-Build/118//console
This message is automatically generated.
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Thomas Graves (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13482368#comment-13482368 ]
Thomas Graves commented on YARN-177:
------------------------------------
well I guess I didn't fix it. The issue is we are inconsistent about synchronizing around the parent usage.
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13481875#comment-13481875 ]
Hadoop QA commented on YARN-177:
--------------------------------
{color:red}-1 overall{color}. Here are the results of testing the latest attachment
http://issues.apache.org/jira/secure/attachment/12550363/YARN-177.patch
against trunk revision .
{color:green}+1 @author{color}. The patch does not contain any @author tags.
{color:red}-1 tests included{color}. The patch doesn't appear to include any new or modified tests.
Please justify why no new tests are needed for this patch.
Also please list what manual steps were performed to verify this patch.
{color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings.
{color:green}+1 javadoc{color}. The javadoc tool did not generate any warning messages.
{color:green}+1 eclipse:eclipse{color}. The patch built with eclipse:eclipse.
{color:red}-1 findbugs{color}. The patch appears to introduce 1 new Findbugs (version 1.3.9) warnings.
{color:green}+1 release audit{color}. The applied patch does not increase the total number of release audit warnings.
{color:green}+1 core tests{color}. The patch passed unit tests in hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager.
{color:green}+1 contrib tests{color}. The patch passed contrib unit tests.
Test results: https://builds.apache.org/job/PreCommit-YARN-Build/112//testReport/
Findbugs warnings: https://builds.apache.org/job/PreCommit-YARN-Build/112//artifact/trunk/patchprocess/newPatchFindbugsWarningshadoop-yarn-server-resourcemanager.html
Console output: https://builds.apache.org/job/PreCommit-YARN-Build/112//console
This message is automatically generated.
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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] [Assigned] (YARN-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Arun C Murthy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Arun C Murthy reassigned YARN-177:
----------------------------------
Assignee: Arun C Murthy (was: Thomas Graves)
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Thomas Graves (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13483211#comment-13483211 ]
Thomas Graves commented on YARN-177:
------------------------------------
+1. Thanks Arun! I'll commit this shortly.
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch, YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13482563#comment-13482563 ]
Hadoop QA commented on YARN-177:
--------------------------------
{color:green}+1 overall{color}. Here are the results of testing the latest attachment
http://issues.apache.org/jira/secure/attachment/12550506/YARN-177.patch
against trunk revision .
{color:green}+1 @author{color}. The patch does not contain any @author tags.
{color:green}+1 tests included{color}. The patch appears to include 1 new or modified test files.
{color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings.
{color:green}+1 javadoc{color}. The javadoc tool did not generate any warning messages.
{color:green}+1 eclipse:eclipse{color}. The patch built with eclipse:eclipse.
{color:green}+1 findbugs{color}. The patch does not introduce any new Findbugs (version 1.3.9) warnings.
{color:green}+1 release audit{color}. The applied patch does not increase the total number of release audit warnings.
{color:green}+1 core tests{color}. The patch passed unit tests in hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager.
{color:green}+1 contrib tests{color}. The patch passed contrib unit tests.
Test results: https://builds.apache.org/job/PreCommit-YARN-Build/119//testReport/
Console output: https://builds.apache.org/job/PreCommit-YARN-Build/119//console
This message is automatically generated.
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch, YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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] [Updated] (YARN-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Arun C Murthy (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Arun C Murthy updated YARN-177:
-------------------------------
Attachment: YARN-177.patch
Fixed findbugs and added a unit test.
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Hudson (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13483221#comment-13483221 ]
Hudson commented on YARN-177:
-----------------------------
Integrated in Hadoop-trunk-Commit #2920 (See [https://builds.apache.org/job/Hadoop-trunk-Commit/2920/])
YARN-177. CapacityScheduler - adding a queue while the RM is running has wacky results (acmurthy vai tgraves) (Revision 1401668)
Result = SUCCESS
tgraves : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1401668
Files :
* /hadoop/common/trunk/hadoop-yarn-project/CHANGES.txt
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CSQueue.java
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/ParentQueue.java
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestCapacityScheduler.java
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch, YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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] [Updated] (YARN-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Thomas Graves (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Thomas Graves updated YARN-177:
-------------------------------
Attachment: YARN-177.patch
Thanks Arun. Here is an updated patch fixing the findbugs warning. Looks good to me other then adding some unit tests like you said.
I manually tested with both adding some multi-level queues and a single level queue and am no longer seeing the issue.
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Thomas Graves (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13482397#comment-13482397 ]
Thomas Graves commented on YARN-177:
------------------------------------
The LeafQueue has a setParentQueue that is unused and can be removed now.
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Hudson (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13484037#comment-13484037 ]
Hudson commented on YARN-177:
-----------------------------
Integrated in Hadoop-Yarn-trunk #16 (See [https://builds.apache.org/job/Hadoop-Yarn-trunk/16/])
YARN-177. CapacityScheduler - adding a queue while the RM is running has wacky results (acmurthy vai tgraves) (Revision 1401668)
Result = SUCCESS
tgraves : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1401668
Files :
* /hadoop/common/trunk/hadoop-yarn-project/CHANGES.txt
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CSQueue.java
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/ParentQueue.java
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestCapacityScheduler.java
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 3.0.0, 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch, YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Hadoop QA (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13482353#comment-13482353 ]
Hadoop QA commented on YARN-177:
--------------------------------
{color:red}-1 overall{color}. Here are the results of testing the latest attachment
http://issues.apache.org/jira/secure/attachment/12550459/YARN-177.patch
against trunk revision .
{color:green}+1 @author{color}. The patch does not contain any @author tags.
{color:red}-1 tests included{color}. The patch doesn't appear to include any new or modified tests.
Please justify why no new tests are needed for this patch.
Also please list what manual steps were performed to verify this patch.
{color:green}+1 javac{color}. The applied patch does not increase the total number of javac compiler warnings.
{color:green}+1 javadoc{color}. The javadoc tool did not generate any warning messages.
{color:green}+1 eclipse:eclipse{color}. The patch built with eclipse:eclipse.
{color:red}-1 findbugs{color}. The patch appears to introduce 1 new Findbugs (version 1.3.9) warnings.
{color:green}+1 release audit{color}. The applied patch does not increase the total number of release audit warnings.
{color:green}+1 core tests{color}. The patch passed unit tests in hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager.
{color:green}+1 contrib tests{color}. The patch passed contrib unit tests.
Test results: https://builds.apache.org/job/PreCommit-YARN-Build/117//testReport/
Findbugs warnings: https://builds.apache.org/job/PreCommit-YARN-Build/117//artifact/trunk/patchprocess/newPatchFindbugsWarningshadoop-yarn-server-resourcemanager.html
Console output: https://builds.apache.org/job/PreCommit-YARN-Build/117//console
This message is automatically generated.
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Hudson (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13484137#comment-13484137 ]
Hudson commented on YARN-177:
-----------------------------
Integrated in Hadoop-Mapreduce-trunk #1236 (See [https://builds.apache.org/job/Hadoop-Mapreduce-trunk/1236/])
YARN-177. CapacityScheduler - adding a queue while the RM is running has wacky results (acmurthy vai tgraves) (Revision 1401668)
Result = FAILURE
tgraves : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1401668
Files :
* /hadoop/common/trunk/hadoop-yarn-project/CHANGES.txt
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CSQueue.java
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/ParentQueue.java
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestCapacityScheduler.java
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 3.0.0, 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch, YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "nemon lou (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13481243#comment-13481243 ]
nemon lou commented on YARN-177:
--------------------------------
I met the same issue in hadoop 2.0.1
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Thomas Graves
> Priority: Critical
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Hudson (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13484096#comment-13484096 ]
Hudson commented on YARN-177:
-----------------------------
Integrated in Hadoop-Hdfs-trunk #1206 (See [https://builds.apache.org/job/Hadoop-Hdfs-trunk/1206/])
YARN-177. CapacityScheduler - adding a queue while the RM is running has wacky results (acmurthy vai tgraves) (Revision 1401668)
Result = SUCCESS
tgraves : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1401668
Files :
* /hadoop/common/trunk/hadoop-yarn-project/CHANGES.txt
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CSQueue.java
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/ParentQueue.java
* /hadoop/common/trunk/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestCapacityScheduler.java
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 3.0.0, 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch, YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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-177) CapacityScheduler - adding a queue
while the RM is running has wacky results
Posted by "Hudson (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/YARN-177?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13484080#comment-13484080 ]
Hudson commented on YARN-177:
-----------------------------
Integrated in Hadoop-Hdfs-0.23-Build #415 (See [https://builds.apache.org/job/Hadoop-Hdfs-0.23-Build/415/])
YARN-177. CapacityScheduler - adding a queue while the RM is running has wacky results (acmurthy vai tgraves) (Revision 1401699)
Result = SUCCESS
tgraves : http://svn.apache.org/viewcvs.cgi/?root=Apache-SVN&view=rev&rev=1401699
Files :
* /hadoop/common/branches/branch-0.23/hadoop-yarn-project/CHANGES.txt
* /hadoop/common/branches/branch-0.23/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/CSQueue.java
* /hadoop/common/branches/branch-0.23/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/LeafQueue.java
* /hadoop/common/branches/branch-0.23/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/main/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/ParentQueue.java
* /hadoop/common/branches/branch-0.23/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/src/test/java/org/apache/hadoop/yarn/server/resourcemanager/scheduler/capacity/TestCapacityScheduler.java
> CapacityScheduler - adding a queue while the RM is running has wacky results
> ----------------------------------------------------------------------------
>
> Key: YARN-177
> URL: https://issues.apache.org/jira/browse/YARN-177
> Project: Hadoop YARN
> Issue Type: Bug
> Components: capacityscheduler
> Affects Versions: 0.23.3
> Reporter: Thomas Graves
> Assignee: Arun C Murthy
> Priority: Critical
> Fix For: 3.0.0, 2.0.3-alpha, 0.23.5
>
> Attachments: YARN-177.patch, YARN-177.patch, YARN-177.patch, YARN-177.patch
>
>
> Adding a queue to the capacity scheduler while the RM is running and then running a job in the queue added results in very strange behavior. The cluster Total Memory can either decrease or increase. We had a cluster where total memory decreased to almost 1/6th the capacity. Running on a small test cluster resulted in the capacity going up by simply adding a queue and running wordcount.
> Looking at the RM logs, used memory can go negative but other logs show the number positive:
> 2012-10-21 22:56:44,796 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: assignedContainer queue=root usedCapacity=0.0375 absoluteUsedCapacity=0.0375 used=memory: 7680 cluster=memory: 204800
> 2012-10-21 22:56:45,831 [ResourceManager Event Processor] INFO org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.ParentQueue: completedContainer queue=root usedCapacity=-0.0225 absoluteUsedCapacity=-0.0225 used=memory: -4608 cluster=memory: 204800
>
--
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