tajo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jihoon Son (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (TAJO-298) Catalog Federation
Date Mon, 04 Nov 2013 12:10:22 GMT

    [ https://issues.apache.org/jira/browse/TAJO-298?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13812792#comment-13812792

Jihoon Son commented on TAJO-298:

If Tajo supports various catalog instances simultaneously, it will be useful.
But I think that it will significantly increase the system's complexity.
For example, when a worker tries to access to a catalog to build a query plan, it should get
the information of which catalog stores the meta information of the relations relevant to
the query plan.
I think that the complexity problem outweighs the benefits.

HDFS federation is designed and implemented to achieve the scalability and to handle the single
point of failures. 
Since the master and workers access the catalog in Tajo, we may need to design the Catalog
federation as the same reason of HDFS federation.

> Catalog Federation
> ------------------
>                 Key: TAJO-298
>                 URL: https://issues.apache.org/jira/browse/TAJO-298
>             Project: Tajo
>          Issue Type: Improvement
>          Components: catalog
>    Affects Versions: 0.8-incubating
>            Reporter: JaeHwa Jung
>            Assignee: JaeHwa Jung
>             Fix For: 0.8-incubating
> Current Catalog supports just one Catalog server. But I think that users want to use
serveral tables stored of multiple Catalogs at the same time. For example, After user query
hive tables, user can inserver into tajo table. Above this, many use cases will happen. So,
I wish that tajo supports muliple Catalog named Catalog Ferderation. I was inspired from HDFS

This message was sent by Atlassian JIRA

View raw message