[
https://issues.apache.org/jira/browse/HBASE-5231?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13192452#comment-13192452
]
Zhihong Yu commented on HBASE-5231:
-----------------------------------
bq. There is no test to prove it
I confine the impact for this feature within the limit of balancer slop which
was introduced more than 6 months ago. This means there is no real change in
the overall balancing for the whole cluster.
The addendum adds javadoc to getAssignmentsByTable().
Part of the reason for putting some of balance-by-table logic in Master is to
limit the complexity of DefaultLoadBalancer.
There may be better ways to simplify the logic in DefaultLoadBalancer. I will
create a separate JIRA for that.
One intrinsic limitation of regions per regionserver being the sole balancing
criterion is that the actual load on the regions is not taken into account.
I have work pending open-source which addresses this limitation.
> Backport HBASE-3373 (per-table load balancing) to 0.92
> ------------------------------------------------------
>
> Key: HBASE-5231
> URL: https://issues.apache.org/jira/browse/HBASE-5231
> Project: HBase
> Issue Type: Improvement
> Reporter: Zhihong Yu
> Fix For: 0.92.1
>
> Attachments: 5231-v2.txt, 5231.addendum, 5231.txt
>
>
> This JIRA backports per-table load balancing to 0.90
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
