[jira] [Commented] (HBASE-12007) StochasticBalancer should avoid putting user regions on master

2014-09-19 Thread Hudson (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-12007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14141225#comment-14141225
 ] 

Hudson commented on HBASE-12007:


FAILURE: Integrated in HBase-TRUNK #5533 (See 
[https://builds.apache.org/job/HBase-TRUNK/5533/])
HBASE-12007 StochasticBalancer should avoid putting user regions on master 
(jxiang: rev 74c6b33e84df7fd45875669f401400f02c004dfa)
* 
hbase-server/src/main/java/org/apache/hadoop/hbase/master/balancer/BaseLoadBalancer.java
* hbase-server/src/test/java/org/apache/hadoop/hbase/TestRegionRebalancing.java
* 
hbase-server/src/test/java/org/apache/hadoop/hbase/master/balancer/TestStochasticLoadBalancer.java
* 
hbase-server/src/main/java/org/apache/hadoop/hbase/master/balancer/StochasticLoadBalancer.java


> StochasticBalancer should avoid putting user regions on master
> --
>
> Key: HBASE-12007
> URL: https://issues.apache.org/jira/browse/HBASE-12007
> Project: HBase
>  Issue Type: Bug
>Reporter: Jimmy Xiang
>Assignee: Jimmy Xiang
> Fix For: 2.0.0, 0.99.1
>
> Attachments: hbase-12007.patch
>
>
> We should enhance how StochasticBalancer picks up a random server so that it 
> can avoid putting user regions on master, because regions on master are 
> handled differently in a separate method.



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


[jira] [Commented] (HBASE-12007) StochasticBalancer should avoid putting user regions on master

2014-09-19 Thread Hudson (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-12007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14141171#comment-14141171
 ] 

Hudson commented on HBASE-12007:


FAILURE: Integrated in HBase-1.0 #207 (See 
[https://builds.apache.org/job/HBase-1.0/207/])
HBASE-12007 StochasticBalancer should avoid putting user regions on master 
(jxiang: rev 57f6ff50893583ed7c8938c50eb67b363fbe7f30)
* 
hbase-server/src/test/java/org/apache/hadoop/hbase/master/balancer/TestStochasticLoadBalancer.java
* hbase-server/src/test/java/org/apache/hadoop/hbase/TestRegionRebalancing.java
* 
hbase-server/src/main/java/org/apache/hadoop/hbase/master/balancer/StochasticLoadBalancer.java
* 
hbase-server/src/main/java/org/apache/hadoop/hbase/master/balancer/BaseLoadBalancer.java


> StochasticBalancer should avoid putting user regions on master
> --
>
> Key: HBASE-12007
> URL: https://issues.apache.org/jira/browse/HBASE-12007
> Project: HBase
>  Issue Type: Bug
>Reporter: Jimmy Xiang
>Assignee: Jimmy Xiang
> Fix For: 2.0.0, 0.99.1
>
> Attachments: hbase-12007.patch
>
>
> We should enhance how StochasticBalancer picks up a random server so that it 
> can avoid putting user regions on master, because regions on master are 
> handled differently in a separate method.



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


[jira] [Commented] (HBASE-12007) StochasticBalancer should avoid putting user regions on master

2014-09-19 Thread Elliott Clark (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-12007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14141069#comment-14141069
 ] 

Elliott Clark commented on HBASE-12007:
---

+1 It's too bad that there is a split on how we balance masters and normal 
regionservers. We should think about healing that divide sometime in the future.

> StochasticBalancer should avoid putting user regions on master
> --
>
> Key: HBASE-12007
> URL: https://issues.apache.org/jira/browse/HBASE-12007
> Project: HBase
>  Issue Type: Bug
>Reporter: Jimmy Xiang
>Assignee: Jimmy Xiang
> Attachments: hbase-12007.patch
>
>
> We should enhance how StochasticBalancer picks up a random server so that it 
> can avoid putting user regions on master, because regions on master are 
> handled differently in a separate method.



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


[jira] [Commented] (HBASE-12007) StochasticBalancer should avoid putting user regions on master

2014-09-19 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-12007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14141033#comment-14141033
 ] 

stack commented on HBASE-12007:
---

+1 on patch then [~jxiang] LGTM

> StochasticBalancer should avoid putting user regions on master
> --
>
> Key: HBASE-12007
> URL: https://issues.apache.org/jira/browse/HBASE-12007
> Project: HBase
>  Issue Type: Bug
>Reporter: Jimmy Xiang
>Assignee: Jimmy Xiang
> Attachments: hbase-12007.patch
>
>
> We should enhance how StochasticBalancer picks up a random server so that it 
> can avoid putting user regions on master, because regions on master are 
> handled differently in a separate method.



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


[jira] [Commented] (HBASE-12007) StochasticBalancer should avoid putting user regions on master

2014-09-19 Thread Jimmy Xiang (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-12007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14141018#comment-14141018
 ] 

Jimmy Xiang commented on HBASE-12007:
-

Yes, it is off by default in branch-1. Only one way of distributing regions.

> StochasticBalancer should avoid putting user regions on master
> --
>
> Key: HBASE-12007
> URL: https://issues.apache.org/jira/browse/HBASE-12007
> Project: HBase
>  Issue Type: Bug
>Reporter: Jimmy Xiang
>Assignee: Jimmy Xiang
> Attachments: hbase-12007.patch
>
>
> We should enhance how StochasticBalancer picks up a random server so that it 
> can avoid putting user regions on master, because regions on master are 
> handled differently in a separate method.



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


[jira] [Commented] (HBASE-12007) StochasticBalancer should avoid putting user regions on master

2014-09-19 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-12007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14140951#comment-14140951
 ] 

stack commented on HBASE-12007:
---

bq. We could assign to backup master?

But it is off by default in branch-1?

bq. Very hard, and it is useful for most of use cases.

But it is off in branch-1 (smile).  I think it fine that M can be a RS but that 
it is not on by default.

bq. Regions on master are handled with balanceMasterRegions which is internal 
to the balancer.
... ok, so only one way of distributing regions... the balancer?

Thanks for bringing me along [~jxiang]



> StochasticBalancer should avoid putting user regions on master
> --
>
> Key: HBASE-12007
> URL: https://issues.apache.org/jira/browse/HBASE-12007
> Project: HBase
>  Issue Type: Bug
>Reporter: Jimmy Xiang
>Assignee: Jimmy Xiang
> Attachments: hbase-12007.patch
>
>
> We should enhance how StochasticBalancer picks up a random server so that it 
> can avoid putting user regions on master, because regions on master are 
> handled differently in a separate method.



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


[jira] [Commented] (HBASE-12007) StochasticBalancer should avoid putting user regions on master

2014-09-19 Thread Jimmy Xiang (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-12007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14140753#comment-14140753
 ] 

Jimmy Xiang commented on HBASE-12007:
-

bq. We have two mechanisms for distributing regions on a cluster?
Regions on master are handled with balanceMasterRegions which is internal to 
the balancer.
bq. In branch-1, master can carry regions? DIdn't we disable that for branch-1?
Yes, it still can, if configured. It doesn't host any region by default.
bq. Or is it not possible to undo M being a RS?
Very hard, and it is useful for most of use cases.
bq. We could assign to backup master?
Yes, that's right.

> StochasticBalancer should avoid putting user regions on master
> --
>
> Key: HBASE-12007
> URL: https://issues.apache.org/jira/browse/HBASE-12007
> Project: HBase
>  Issue Type: Bug
>Reporter: Jimmy Xiang
>Assignee: Jimmy Xiang
> Attachments: hbase-12007.patch
>
>
> We should enhance how StochasticBalancer picks up a random server so that it 
> can avoid putting user regions on master, because regions on master are 
> handled differently in a separate method.



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


[jira] [Commented] (HBASE-12007) StochasticBalancer should avoid putting user regions on master

2014-09-18 Thread stack (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-12007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14139801#comment-14139801
 ] 

stack commented on HBASE-12007:
---

bq. The regions in the active master are already handled separately.

When you say the above, what does that mean? We have two mechanisms for 
distributing regions on a cluster?

In branch-1, master can carry regions?  DIdn't we disable that for branch-1?  
Or is it not possible to undo M being a RS?

bq. ...&& cluster.isActiveMaster(n)

We could assign to backup master?

Thanks

> StochasticBalancer should avoid putting user regions on master
> --
>
> Key: HBASE-12007
> URL: https://issues.apache.org/jira/browse/HBASE-12007
> Project: HBase
>  Issue Type: Bug
>Reporter: Jimmy Xiang
>Assignee: Jimmy Xiang
> Attachments: hbase-12007.patch
>
>
> We should enhance how StochasticBalancer picks up a random server so that it 
> can avoid putting user regions on master, because regions on master are 
> handled differently in a separate method.



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


[jira] [Commented] (HBASE-12007) StochasticBalancer should avoid putting user regions on master

2014-09-18 Thread Jimmy Xiang (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-12007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14139151#comment-14139151
 ] 

Jimmy Xiang commented on HBASE-12007:
-

I meant any region that is assigned to the master. In branch-1, 
StochasticBalancer could assign some region to the master since there is no 
region on it, per the region count skew cost function, then the balancer will 
move the region away in the another run, which causes the flakiness of 
TestRegionRebalancing. With this patch, we leave balancing regions on the 
master server to the base balancer instead of StochasticBalancer. 

> StochasticBalancer should avoid putting user regions on master
> --
>
> Key: HBASE-12007
> URL: https://issues.apache.org/jira/browse/HBASE-12007
> Project: HBase
>  Issue Type: Bug
>Reporter: Jimmy Xiang
>Assignee: Jimmy Xiang
> Attachments: hbase-12007.patch
>
>
> We should enhance how StochasticBalancer picks up a random server so that it 
> can avoid putting user regions on master, because regions on master are 
> handled differently in a separate method.



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


[jira] [Commented] (HBASE-12007) StochasticBalancer should avoid putting user regions on master

2014-09-18 Thread Sergey Soldatov (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-12007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14138870#comment-14138870
 ] 

Sergey Soldatov commented on HBASE-12007:
-

When you say about regions on master server do you mean meta, acl and namespace 
regions? In branch-1 they are not assigned to master.

> StochasticBalancer should avoid putting user regions on master
> --
>
> Key: HBASE-12007
> URL: https://issues.apache.org/jira/browse/HBASE-12007
> Project: HBase
>  Issue Type: Bug
>Reporter: Jimmy Xiang
>Assignee: Jimmy Xiang
> Attachments: hbase-12007.patch
>
>
> We should enhance how StochasticBalancer picks up a random server so that it 
> can avoid putting user regions on master, because regions on master are 
> handled differently in a separate method.



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


[jira] [Commented] (HBASE-12007) StochasticBalancer should avoid putting user regions on master

2014-09-17 Thread Hadoop QA (JIRA)

[ 
https://issues.apache.org/jira/browse/HBASE-12007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14138113#comment-14138113
 ] 

Hadoop QA commented on HBASE-12007:
---

{color:red}-1 overall{color}.  Here are the results of testing the latest 
attachment 
  http://issues.apache.org/jira/secure/attachment/12669523/hbase-12007.patch
  against trunk revision .
  ATTACHMENT ID: 12669523

{color:green}+1 @author{color}.  The patch does not contain any @author 
tags.

{color:green}+1 tests included{color}.  The patch appears to include 6 new 
or modified tests.

{color:green}+1 javac{color}.  The applied patch does not increase the 
total number of javac compiler warnings.

{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 findbugs{color}.  The patch does not introduce any new 
Findbugs (version 2.0.3) warnings.

{color:green}+1 release audit{color}.  The applied patch does not increase 
the total number of release audit warnings.

{color:green}+1 lineLengths{color}.  The patch does not introduce lines 
longer than 100

  {color:green}+1 site{color}.  The mvn site goal succeeds with this patch.

 {color:red}-1 core tests{color}.  The patch failed these unit tests:
 

Test results: 
https://builds.apache.org/job/PreCommit-HBASE-Build/10953//testReport/
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/10953//artifact/patchprocess/newPatchFindbugsWarningshbase-hadoop2-compat.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/10953//artifact/patchprocess/newPatchFindbugsWarningshbase-prefix-tree.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/10953//artifact/patchprocess/newPatchFindbugsWarningshbase-common.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/10953//artifact/patchprocess/newPatchFindbugsWarningshbase-thrift.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/10953//artifact/patchprocess/newPatchFindbugsWarningshbase-examples.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/10953//artifact/patchprocess/newPatchFindbugsWarningshbase-client.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/10953//artifact/patchprocess/newPatchFindbugsWarningshbase-hadoop-compat.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/10953//artifact/patchprocess/newPatchFindbugsWarningshbase-server.html
Findbugs warnings: 
https://builds.apache.org/job/PreCommit-HBASE-Build/10953//artifact/patchprocess/newPatchFindbugsWarningshbase-protocol.html
Console output: 
https://builds.apache.org/job/PreCommit-HBASE-Build/10953//console

This message is automatically generated.

> StochasticBalancer should avoid putting user regions on master
> --
>
> Key: HBASE-12007
> URL: https://issues.apache.org/jira/browse/HBASE-12007
> Project: HBase
>  Issue Type: Bug
>Reporter: Jimmy Xiang
>Assignee: Jimmy Xiang
> Attachments: hbase-12007.patch
>
>
> We should enhance how StochasticBalancer picks up a random server so that it 
> can avoid putting user regions on master, because regions on master are 
> handled differently in a separate method.



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