You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Nick Dimiduk (Jira)" <ji...@apache.org> on 2020/06/25 22:02:00 UTC
[jira] [Commented] (HBASE-21782) LoadIncrementalHFiles should not
be IA.Public
[ https://issues.apache.org/jira/browse/HBASE-21782?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17145885#comment-17145885 ]
Nick Dimiduk commented on HBASE-21782:
--------------------------------------
[~zhangduo] the deprecation comments from this change on {{o.a.h.h.tool.LoadIncrementalHFiles}} say "deprecated since 2.2.0, removal in 3.0.0", which does not meet our guideline re: deprecation for an entire major version. I believe we cannot drop this class until 4.0.
> LoadIncrementalHFiles should not be IA.Public
> ---------------------------------------------
>
> Key: HBASE-21782
> URL: https://issues.apache.org/jira/browse/HBASE-21782
> Project: HBase
> Issue Type: Task
> Components: mapreduce
> Reporter: Duo Zhang
> Assignee: Duo Zhang
> Priority: Major
> Labels: bulkload
> Fix For: 3.0.0-alpha-1, 2.2.0
>
> Attachments: HBASE-21782-v1.patch, HBASE-21782.patch
>
>
> It is an implementation class, so some of the methods which are only supposed to be used by replication sink are also public to users. And it exposes methods which take Table and Connection as parameter and inside the implementation we assume that they are HTable and ConnectionImplementation, which will be a pain when we want to replace the sync client implementation with async client.
> Here I think we should make the implementation class as IA.LimitPrivate(TOOL), and introduce an interface for bulking hfiles programmatically.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)