[
https://issues.apache.org/jira/browse/HBASE-16040?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15333769#comment-15333769
]
Sean Busbey commented on HBASE-16040:
-------------------------------------
I think the main purpose this currently serves is a single place one could
toggle on/off replication to allow WALs to be cleared without adding/removing
replication peers. Can anyone else think of a different use?
In practice, I've always just documented the info on replication peers and then
removed/added them back to handle this kind of issue. a rolling restart of the
cluster just seemed a higher cost.
> Remove configuration "hbase.replication"
> ----------------------------------------
>
> Key: HBASE-16040
> URL: https://issues.apache.org/jira/browse/HBASE-16040
> Project: HBase
> Issue Type: Bug
> Reporter: Heng Chen
> Fix For: 2.0.0
>
>
> This configuration was introduced to reduce overhead of replication.
> Now the overhead of replication is negligible. Besides that, this config is
> not in hbase-default.xml, user has to read the code to know about it and
> its' default value, this is unfriendly.
> So let's remove it. suggestions?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
