[
https://issues.apache.org/jira/browse/HBASE-25293?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Huaxiang Sun resolved HBASE-25293.
----------------------------------
Fix Version/s: 2.4.1
Resolution: Fixed
> Followup jira to address the client handling issue when chaning from meta
> replica to non-meta-replica at the server side.
> -------------------------------------------------------------------------------------------------------------------------
>
> Key: HBASE-25293
> URL: https://issues.apache.org/jira/browse/HBASE-25293
> Project: HBase
> Issue Type: Sub-task
> Reporter: Huaxiang Sun
> Assignee: Huaxiang Sun
> Priority: Minor
> Fix For: 2.4.1
>
>
> [https://github.com/apache/hbase/pull/2643]
>
> {quote}
> With my operator hat on, I'd assume that LOAD_BALANCE with 1 replica count
> works like no read replicas configured (logic wise at-least, even though the
> code paths are different).
> {quote}If the server side does not support meta replica, the client side
> cannot be configured to support this mode
> {quote}
> Since clients are usually long running (meaning we may not be able to restart
> client or they using cached HBase connection) and meta replica count can be
> altered on the service side on the fly, I'd expect client to work across
> these changes without any configuration changes. WDYT?
> {quote}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
