You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hbase.apache.org by "Andrew Purtell (JIRA)" <ji...@apache.org> on 2014/07/19 02:22:40 UTC
[jira] [Updated] (HBASE-3250) Clean up HBaseAdmin APIs
[ https://issues.apache.org/jira/browse/HBASE-3250?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Andrew Purtell updated HBASE-3250:
----------------------------------
Affects Version/s: (was: 0.92.0)
Fix Version/s: 1.0.0
Revisit for 1.0?
> Clean up HBaseAdmin APIs
> ------------------------
>
> Key: HBASE-3250
> URL: https://issues.apache.org/jira/browse/HBASE-3250
> Project: HBase
> Issue Type: Improvement
> Reporter: Todd Lipcon
> Fix For: 1.0.0
>
>
> Some discussion on IRC about this - HBaseAdmin is a bit of a mess currently, it has a lot of different calls. They tend to fall into these categories:
> - Things that actually affect the data from a user perspective (eg adding/dropping/enable/disable tables, setting up CFs, etc)
> - Things that affect the underlying data storage (eg force split, flush, compact)
> - Cluster status (isMasterRunning, checkHBaseAvailable)
> I'd propose we separate these into different classes. HBaseAdmin would be reserved for "DDL" operations (eg create/alter/drop table). A new class, eg HBaseRegionTools or HBaseStorageAdmin or something, would handle the calls that affect "implementation details" like split/flush/compact. The cluster status stuff should maybe just go in a new HBaseCluster class which would also have getAdmin(), getTable(), etc?
--
This message was sent by Atlassian JIRA
(v6.2#6252)