[
https://issues.apache.org/jira/browse/CHUKWA-490?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Eric Yang updated CHUKWA-490:
-----------------------------
Status: Resolved (was: Patch Available)
Resolution: Fixed
I just committed this, thanks Kirk.
> Allow mdl.xml to be split into multiple files for ease of maintenance
> ---------------------------------------------------------------------
>
> Key: CHUKWA-490
> URL: https://issues.apache.org/jira/browse/CHUKWA-490
> Project: Hadoop Chukwa
> Issue Type: Improvement
> Affects Versions: 0.4.0
> Reporter: Kirk True
> Attachments: DatabaseConfig.diff
>
>
> We have nearly two dozen different demux implementations that generate data
> that is eventually saved into MySQL. On top of this, each different data type
> has about twenty or more columns in its resulting table. Specifying all of
> these in a single mdl.xml file has caused us some headaches, so we looked
> into how we could split mdl.xml into smaller, more manageable chunks. The
> patch (forthcoming) simply iterates over the files in $CHUKWA_CONF_DIR and if
> the file name ends with "mdl.xml", it will be added as a resource to the
> Configuration object used (mdl.xml is loaded first and the others are loaded
> subsequently, ensuring that mdl.xml is not then loaded again).
> One likely concern is that this doesn't follow the default/site method of
> specifying overrides. Also, the ordering of the configuration files'
> inclusionis based on the JVM and/or platform implementation of
> java.io.File.listFiles.
> Regardless, this made it very simple to isolate the files needed by each data
> type and update them as needed.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
