incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Eric Sammer <esam...@cloudera.com>
Subject Re: [PROPOSAL] MRUnit entry into the Apache Incubator
Date Tue, 01 Mar 2011 17:22:16 GMT
On Mon, Feb 28, 2011 at 10:18 AM, ant elder <ant.elder@gmail.com> wrote:

> On Mon, Feb 28, 2011 at 2:57 PM, Mattmann, Chris A (388J)
> <chris.a.mattmann@jpl.nasa.gov> wrote:
> > Hi Guys,
> >
> >>>
> >>> I don't know, none of those things really sound like justification for
> >>> a separate TLP, keeping something like this nearby to Hadoop MapReduce
> >>> seems like it would help give it a better chance of long term survival
> >>>  to me. Wouldn't simply moving mrunit out from
> >>> hadoop/mapreduce/trunk/src to somewhere else so it can do its own
> >>> releases and deploy its artifacts achieve all those objectives?...
> >
> > Personally I don't believe that the project has any chance for survival
> at Apache unless this Incubator project happens.
> >
> > The Hadoop community has been VOTE'ing over the past month to abandon
> development of portions of their code base (namely contrib portions that no
> developers or users are maintaining or leveraging). MRUnit came up as one of
> those contribs, and what was discovered was that some of the issues with the
> Hadoop community (electing new committers, having too many products tied to
> a single release cycle, too many infrequent releases) had stagnated the
> development of MRUnit so much that a small sect of Hadoop developers had
> forked MRUnit into Github and were developing it there. So there *was*
> active development going on, it just wasn't at Apache.
> >
> > A few of us members (myself, Patrick and Nige) were willing to step up
> and volunteer that in contrast to seeing an ASF product and its community
> head over to Github and continue development there, we'd like the product to
> stick around at the ASF.
> >
> > There is no hard and fast rule as to the size of a community for
> TLP-status, and that's far down the road anyways aka why are we worrying
> about that now during a [PROPOSAL] phase when it's an *exit criteria* for
> the Incubator? The point of Incubation is to help grow that community and
> sort that stuff out and I have no doubt in my mind that these folks here on
> the proposal who are willing and trying to develop something successful at
> Apache can't meet that challenge.
> >
>
> You're right about the entry criteria, there's a very low bar to
> becoming an Incubator poddling so if you all really want to I'm pretty
> sure you'll be able to make it happen. But this is a proposal
> discussion thread which asks for feedback so thats what i gave.
>

Definitely appreciated, ant.


> Now that you've explained further I even more think it doesn't seem
> quite right. Sounds to me like the Hadoop PMC is a bit disfunctional
> if they couldn't make these guys committers and pushed them off to
> github and maybe it would be better to just fix that. Or split
> MapReduce out of Hadoop to its own new TLP which you'd think would be
> more willing to embrace MRUnit.
>

To clarify, it's not that the Hadoop PMC pushed MRUnit out, just that
there's a lot of infrastructure and changes that would be required to be
able to release MRUnit independently within the project. Aaron Kimball (one
of the proposed mrunit committers) is also a Hadoop committer. That said,
the current build, test, release infrastructure of Hadoop doesn't easily
support separate release cycles which makes things a bit difficult. I agree
the ideal situation would be for Hadoop to easily support separate artifacts
and release cycles but I don't think that's in the cards in the short term.

With respect to community, it's true that all users of mrunit are Hadoop
users but it's not necessarily true that all Hadoop users care about the
development of Hadoop itself. If Hadoop were an application server, mrunit
would be developer facing public APIs for application authors.

I don't know if mrunit's final destination is TLP status. What I'd like to
do is have it enter the Incubator to develop a community, evolve the
project, and decide the best permanent home, be it TLP, adoption as a
subproject of Hadoop, or something else. My (lay) understanding of the
Incubator is that this is the appropriate forum for this kind of (project)
development. Your points are well taken and make a lot of sense.

Thanks!


>   ...ant
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: general-unsubscribe@incubator.apache.org
> For additional commands, e-mail: general-help@incubator.apache.org
>
>


-- 
Eric Sammer
twitter: esammer
data: www.cloudera.com

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message