[
https://issues.apache.org/jira/browse/HADOOP-1747?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
stack updated HADOOP-1747:
--------------------------
Attachment: hadoop-1747.patch
Here's a fix. Commit message to use is below.
HADOOP-1747 On a cluster, on restart, regions multiply assigned
M src/contrib/hbase/src/java/org/apache/hadoop/hbase/HMaster.java
Removed some empty lines so I can squeeze more code into a screenful.
(assignedRegions): Factored out some code into own methods so
this method is made a bit shorter. Added early returns near
top -- if nothing to assign, etc. -- so less nesting.
Added fix: Instead of iterating over unassignedRegions after
all the loadings have been calculated, instead iterate over
the locally calculated map, regionsToAssign (Otherwise, we
were running over the same territory each time through the
loop and were thus giving out same region multiple times).
(regionsPerServer, assignRegionsToOneServer,
getRegionsToAssign): Added.
> [hbase] On a cluster, on restart, regions multiply assigned
> -----------------------------------------------------------
>
> Key: HADOOP-1747
> URL: https://issues.apache.org/jira/browse/HADOOP-1747
> Project: Hadoop
> Issue Type: Bug
> Components: contrib/hbase
> Reporter: stack
> Assignee: stack
> Priority: Minor
> Fix For: 0.15.0
>
> Attachments: hadoop-1747.patch
>
>
> In a cluster of multiple region servers, on restart, regions are being
> assigned multiple times often to different servers.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
