[jira] [Updated] (BIGTOP-1067) Testing input splits in jobs

["Konstantin Boudnik (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/BIGTOP-1067?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Konstantin Boudnik updated BIGTOP-1067:
---------------------------------------

    Component/s: Tests

> Testing input splits in jobs
> ----------------------------
>
>                 Key: BIGTOP-1067
>                 URL: https://issues.apache.org/jira/browse/BIGTOP-1067
>             Project: Bigtop
>          Issue Type: Test
>          Components: Tests
>    Affects Versions: 0.7.0
>            Reporter: jay vyas
>            Priority: Minor
>             Fix For: backlog
>
>
> One of the things which seem important for serialization frameworks and changes to custom input formats is splitting behaviour.   Should we have a smoke test template that runs jobs with varying input split sizes, confirming that outputs are identical?  Just an idea at the moment but someone with more insight into serialization frameworks and RecordReader/Writer implementations might have a better concept of the usefullness of such smokes.  
> This is a someone open ended JIRA - any thoughts on the issue of testing hadoop's input formats and splits are welcome.   I can try to implement corresponding smokes accordingly. 



--
This message was sent by Atlassian JIRA
(v6.2#6252)