[jira] Commented: (CHUKWA-292) should allow demux reducers in any namespace

["Jerome Boulon (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CHUKWA-292?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12719654#action_12719654 ] 

Jerome Boulon commented on CHUKWA-292:
--------------------------------------

I used to want aliases on reducer side. However I wonder if we really have to make any aggregation on the reducer side. 
By doing so we cannot remove duplicate after demux.  So I may want to force demux to be parsing only per stream and have a viewManager to combine recordTypes. If that the case I could easily build a remove duplicates function on the reducer side.



> should allow demux reducers in any namespace
> --------------------------------------------
>
>                 Key: CHUKWA-292
>                 URL: https://issues.apache.org/jira/browse/CHUKWA-292
>             Project: Hadoop Chukwa
>          Issue Type: Improvement
>          Components: Data Processors
>            Reporter: Ari Rabkin
>   Original Estimate: 1.5h
>  Remaining Estimate: 1.5h
>
> Right now, the Demux Reduce only looks in org.apache.hadoop.chukwa.extraction.demux.processor.reducer to find reduce processors.  This restriction should be dropped.
> Proposal:  Do what we do on the map side, and try a lookup against chukwa-demux-conf to see if we get a match. 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.