[
https://issues.apache.org/jira/browse/HBASE-10513?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13980480#comment-13980480
]
Enis Soztutar commented on HBASE-10513:
---------------------------------------
Thanks Stack. I'll address the comments.
bq. What is 'High Available Reads'. Is that this feature? Its not called
timeline consistent reads or some such?
I think the best description is "read high availability using timeline
consistent reads". I'll clarify.
bq. Do I have to turn this on cluster-wide or by table? If the latter, maybe
caveat your 2-3X memstore usage....
This is per table. Let me clarify that as well. It only kicks in for tables
having REGION_REPLICATION >1.
bq. Ain't stochastic balancer on by default? If so, why mention it in your
configs section? Might confuse?
Yes, SLB is default. Just wanted to make sure that users are not using a
non-default LB. Let me clarify that this may not be needed if using default.
> Provide user documentation for region replicas
> ----------------------------------------------
>
> Key: HBASE-10513
> URL: https://issues.apache.org/jira/browse/HBASE-10513
> Project: HBase
> Issue Type: Sub-task
> Reporter: Enis Soztutar
> Assignee: Enis Soztutar
> Fix For: 0.99.0
>
> Attachments: UserdocumentationforHBASE-10070.pdf
>
>
> We need some documentation for the feature introduced in HBASE-10070.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
