[
https://issues.apache.org/jira/browse/HBASE-6730?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13451067#comment-13451067
]
Elliott Clark commented on HBASE-6730:
--------------------------------------
For a couple of reasons.
* Currently the chores need to be interrupted when the master is being brought
down and the balancer never gets a notification that things are shutting down.
* The balancer only has an instance of MasterServices which doesn't give access
to getClusterStatus and I didn't want to change that interface.
However I agree that having a bunch of chores hanging off of the master just
seems wrong, way too many threads and way too much in the master. I can look
at moving our chore's into something a little easier to remove from the master
if you want, however that would be a little more code to change.
> Enable rolling averages in StochasticLoadBalancer
> --------------------------------------------------
>
> Key: HBASE-6730
> URL: https://issues.apache.org/jira/browse/HBASE-6730
> Project: HBase
> Issue Type: Improvement
> Reporter: Elliott Clark
> Assignee: Elliott Clark
> Fix For: 0.96.0
>
> Attachments: HBASE-6730-1.patch
>
>
> Now that all of the ServerLoad transitions into pb are done. the load
> balancer should get more updates about the current state of the cluster.
> This should be used to allow StochasticLoadBalancer to get rolling averages.
--
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
