[
https://issues.apache.org/jira/browse/PHOENIX-6988?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17752478#comment-17752478
]
Istvan Toth commented on PHOENIX-6988:
--------------------------------------
IMO we should not add observers here.
Our current approach is to group observers in separate coprocessors called
*Observer by functionality, and I see no reason to change that.
My only issue is only with the name.
If we do not plan to add observers, then I'd prefer the more specifc
_PhoenixRegionServerEndpoint_ name instead of the more generic
_PhoenixRegionServerCoprocessor._
> Create new regionserver coprocessor named PhoenixRegionServerCoprocessor in
> Phoenix
> -----------------------------------------------------------------------------------
>
> Key: PHOENIX-6988
> URL: https://issues.apache.org/jira/browse/PHOENIX-6988
> Project: Phoenix
> Issue Type: Sub-task
> Components: core
> Reporter: Rushabh Shah
> Assignee: Rushabh Shah
> Priority: Major
>
> -Rename ReplicationSinkEndpoint to PhoenixRegionServerCoprocessor so that we
> can add all the phoenix regionserver co-proc methods in just 1 class.-
> Earlier I thought to rename ReplicationSinkCompatEndpoint to
> PhoenixRegionServerCoprocessor but then looking at the code, there are some
> complexities. ReplicationSinkEndpoint extends ReplicationSinkCompatEndpoint
> which is defined in hbase-compat module and it is implemented in different
> hbase-compat modules. The new co-proc is phoenix native co-proc and don't
> need any compatibility guarantees. So I created a new regionserver coproc in
> phoenix-core module.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
