Hey Christian,

Couple of options here: I maintain a CDH-compatible release of Crunch for
my customers by creating a cdh-x.y.z branch off of the apache-crunch-x.y.z
branch that gets created for each release, updating the POM to reference
the CDH version I want to use as a base, adding in the CDH repositories and
distributionManagement info, setting the version numbers for the project,
and then running mvn clean deploy to push the code to the repo. I can send
you a diff file so you can see what I have to change and how I wire it up.

The other option is to use Apache Bigtop to manage the dependency
integration across all of your Hadoop components; Bigtop has Crunch as a
component, so it's not too much trouble to manage the releases that way.
Email Roman Shaposhnik ([email protected]) if you want more info about that
approach.

Josh



On Thu, Dec 12, 2013 at 8:35 AM, Christian Tzolov <
[email protected]> wrote:

> Hi guys,
>
> I would like to create Crunch release compatible with the latest Pivotal
> hadoop distribution (2.0.5-alpha).
> The release will be resolved via the Spring's maven repository (
> http://repo.spring.io/libs-release).
>
> I've been thinking of forking the github mirror (
> https://github.com/apache/crunch) and perfrom the releases on the fork.
>
> What do you think would be the easiest way to maintain such releases?
>
> Thanks,
>  Chris
>



-- 
Director of Data Science
Cloudera <http://www.cloudera.com>
Twitter: @josh_wills <http://twitter.com/josh_wills>

Reply via email to