You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@knox.apache.org by "Kevin Minder (JIRA)" <ji...@apache.org> on 2015/04/20 20:34:00 UTC
[jira] [Updated] (KNOX-189) Use directory as source for topology
information instead of single file
[ https://issues.apache.org/jira/browse/KNOX-189?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Kevin Minder updated KNOX-189:
------------------------------
Fix Version/s: (was: 0.6.0)
0.7.0
> Use directory as source for topology information instead of single file
> -----------------------------------------------------------------------
>
> Key: KNOX-189
> URL: https://issues.apache.org/jira/browse/KNOX-189
> Project: Apache Knox
> Issue Type: Improvement
> Components: Server
> Affects Versions: 0.1.0
> Reporter: Kevin Minder
> Fix For: 0.7.0
>
>
> Currently Knox uses a single file to describe the topology of a Hadoop cluster. Add features have been added this has led to force fitting topology configuration into a name/value pair model. In addition some configuration can be extensive and overly complicates the topology file. The suggestion here is to support a topology directory that can contain multiple files. The name of the directory would be used as the cluster name and it must contain a topology.xml file. All other files in the directory would be referenced either implicitly or explicitly via the topology.xml file. This will obviously complicate the book keeping required to determine when a topology needs to be redeployed. I have used JDBM in the past as a way to store metadata about file timestamps. This or something similar might work well here also.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)