[HIVE-4644] Access Multiple HBase clusters' tables simultaneously

• Type: New Feature
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 0.10.0
• Component/s: Database/Schema, HiveServer2, Metastore
• Assignee: Unassigned
• Reporter: Hari Sekhon
• Created: 02/Jun/13 18:25
• Updated: 02/Jun/13 18:28

Description

Right now it doesn't seem possible to simultaneously access multiple HBase clusters in Hive. The best workaround I've been able to come up with in Hive and BeeLine is:

set hbase.zookeeper.quorum=cluster2-zookeeper-host;
select count(*) from hbase2_test;

set hbase.zookeeper.quorum=cluster1-zookeeper-host;
select count(*) from hbase1_test;

But I'm not sure I can use the hbase.zookeeper.quorum override in JDBC in HiveServer2, at least I haven't seen any example code for config overrides.

This workaround is also non-ideal and doesn't allow for simultaneous access to different HBase clusters. So I've tried fiddling with SerDe, Table and Schema level properties to set hbase.zookeeper.quorum overrides on a per table or per schema basis but that didn't work.

I think if the code can be made to respect those properties it would represent the best way of handling this in future, eg:

create database cluster2 WITH DBPROPERTIES ("hbase.zookeeper.quorum"="cluster2-zookeeper-host");

would allow creation of tables in cluster2 database referencing the second HBase cluster.

I've over-simplified here with one zookeeper host in the quorum just for brevity.