hadoop-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Eric Baldeschwieler <eri...@hortonworks.com>
Subject Re: [DISCUSS] Spin out MR, HDFS and YARN as their own TLPs and disband Hadoop umbrella project
Date Wed, 29 Aug 2012 17:42:54 GMT
Hi Tom,

> There are also Hadoop tools like distcp, Hadoop archives, Streaming,
> etc, which should go with MapReduce.

Good point.  I agree.

> The alternative would be to have a Common TLP,
> which we shouldn't necessarily dismiss, since more important than the
> size of the codebase is that there's a community to support the
> codebase, as there certainly is here. 


I guess the question is who would want to be on that project?  I don't think the current bundle
of stuff in common would form a good kernel for a community.  A lack of a coherent community
for common has always been a problem with the project split IMO.  I could see folks deciding
that they were going to build a community around a really good RPC stack, or some other chunk
of common, but frankly I think it it premature to do that.  Proposals welcome of course, but
I think the HDFS folks will want a copy of the RPC stuff in their project and most of the
rest of the stuff in common is too small to merit a project and is more easily handled via
duplication and then sorting it out / dead code elimination.

On Aug 29, 2012, at 10:30 AM, Tom White wrote:

> On Wed, Aug 29, 2012 at 5:31 PM, Arun C Murthy <acm@hortonworks.com> wrote:
>> 
>> On Aug 28, 2012, at 8:50 PM, Alejandro Abdelnur wrote:
>> 
>>> Chris, thanks for initiating the discussion.
>> 
>> Likewise, thanks Chris!
>> 
>>> 
>>> IMO a pre-requisite to this is to figure out how we'll handle the following:
>>> 
>> 
>> 
>> Good points - I'd recommend we keep Common and HDFS in the same project.
> 
> That seems reasonable. The alternative would be to have a Common TLP,
> which we shouldn't necessarily dismiss, since more important than the
> size of the codebase is that there's a community to support the
> codebase, as there certainly is here. Having said that, a Common TLP
> lacks a clear 'mission' since it doesn't offer any standalone
> services. Also, it may diminish in utility over time if pieces are
> moved into HDFS, MapReduce and YARN.
> 
>> Yes, MR/YARN will need some changes in Common occasionally, but core pieces like
RPC have been maintained by HDFS folks over time anyway e.g. move to ProtoBufs were led by
Sanjay, Suresh, Todd, Jitendra et al.
> 
> Does the work to use versioned protocol buffers for RPC mean that
> different releases of HDFS and MapReduce can work together yet? If
> not, this is something we should be working towards (although that
> shouldn't block a move to TLPs).
> 
>> 
>> We can move SequenceFile into MR if necessary and keep same package names for compatibility.
> 
> There are also Hadoop tools like distcp, Hadoop archives, Streaming,
> etc, which should go with MapReduce.
> 
> Cheers,
> Tom
> 
>> 
>> We should, of course, stop tweaking things in different projects in the same jira
- we've been reasonably good at not doing that.
>> 
>> Thoughts?
>> 
>> Arun
>> 
>>> * Where does common stuff lives?
>>> * What are the public interfaces of each project (towards the other projects)?
>>> * How do we do development/releases? In tandem? Separate? How this
>>> will work in practice, currently we are constantly tweaking things
>>> inter-projects, sometimes in the same JIRAs, sometimes in follow up
>>> JIRAs.
>>> 
>>> Thoughts?
>>> 
>>> Thxs.
>>> 
>>> On Tue, Aug 28, 2012 at 7:33 PM, Mattmann, Chris A (388J)
>>> <chris.a.mattmann@jpl.nasa.gov> wrote:
>>>> [decided to minimize traffic and to simply put this in one thread]
>>>> 
>>>> Hi Guys,
>>>> 
>>>> See the recent discussion on these threads:
>>>> 
>>>> YARN as its own Hadoop "sub project": http://s.apache.org/WW1
>>>> Maintain a single committer list for the Hadoop project: http://s.apache.org/Owx
>>>> 
>>>> ...and just pay attention to the Hadoop project over the last 3-4 years.
It's operating
>>>> as a single project, that's masking separate communities that themselves
are really
>>>> separate ASF projects.
>>>> 
>>>> At the ASF, this has been a problem area called "umbrella" projects and over
the years,
>>>> all I've seen from them is wasted bandwidth, artificial barriers and the
inventions of
>>>> new ways to perform process mongering and to reduce the fun in developing
software
>>>> at this fantastic foundation.
>>>> 
>>>> I've talked about umbrella projects enough. We've diverted conversation enough.
>>>> Enough people have tried to act like there is some technical mumbo jumbo
that is
>>>> preventing the eventual act of higher power that I myself hope comes should
these
>>>> discussions prove unfruitful through normal means.
>>>> 
>>>> *these. are. separate. projects.*
>>>> *there.are.not.blocker.issues.from.spinning.out.these.projects.as.their.own.communities*
>>>> 
>>>> In this email: http://s.apache.org/rSm
>>>> 
>>>> And in the 2 subsequent follow ons in that thread, I've outlined a process
that I'll copy
>>>> through below for splitting these projects into their own TLPs:
>>>> 
>>>> -----snip
>>>> Process:
>>>> 
>>>> 0. [DISCUSS] thread for <TLP name> in which you talk about #1 and #2
below, potentially draft resolution too.
>>>> 
>>>> 1. Decide on an initial set of *PMC* members. I urge each new TLP to adopt
PMC==C. See reasons I've
>>>> already discussed.
>>>> 
>>>> 2. Decide on a chair. Try not to VOTE for this explicitly, see if can be
discussed and consensus
>>>> can be reached (just a thought experiment). VOTE if necessary.
>>>> 
>>>> 3. [VOTE] thread for <TLP name>
>>>> 
>>>> 4. Create Project:
>>>> a. paste resolution from #0 to board@ or;
>>>> b. go to general@incubator and start new Incubator project.
>>>> 
>>>> 5. infrastructure set up.
>>>>  MLs moving; new UNIX groups; website setup;
>>>>  SVN setup like this:
>>>> 
>>>> svn copy -m "MR TLP." https://svn.apache.org/repos/asf/hadoop/ https://svn.apache.org/repos/asf/<insert
cool MR name>; or
>>>> svn copy -m "YARN TLP." https://svn.apache.org/repos/asf/hadoop/ https://svn.apache.org/repos/asf/<insert
cool YARN name>; or
>>>> svn copy -m "HDFS TLP." https://svn.apache.org/repos/asf/hadoop/ https://svn.apache.org/repos/asf/<insert
cool HDFS name>
>>>> 
>>>> After all 3 have been created run:
>>>> 
>>>> svn remove -m "Remove Hadoop umbrella TLP. Split into separate projects."
https://svn.apache.org/repos/asf/hadoop
>>>> 
>>>> 6. (TLPs if 4a; Incubator podling if 4b;) proceed, collaborate, operate as
distinct communities, and try to solve the code duplication/dependency
>>>> issues from there.
>>>> 
>>>> 7. If 4b; then graduate as TLP from Incubator.
>>>> 
>>>> -----snip
>>>> 
>>>> So that's my proposal.
>>>> 
>>>> Thanks guys.
>>>> 
>>>> Cheers,
>>>> Chris
>>>> 
>>>> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>>> Chris Mattmann, Ph.D.
>>>> Senior Computer Scientist
>>>> NASA Jet Propulsion Laboratory Pasadena, CA 91109 USA
>>>> Office: 171-266B, Mailstop: 171-246
>>>> Email: chris.a.mattmann@nasa.gov
>>>> WWW:   http://sunset.usc.edu/~mattmann/
>>>> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>>> Adjunct Assistant Professor, Computer Science Department
>>>> University of Southern California, Los Angeles, CA 90089 USA
>>>> ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
>>>> 
>>> 
>>> 
>>> 
>>> --
>>> Alejandro
>> 
>> --
>> Arun C. Murthy
>> Hortonworks Inc.
>> http://hortonworks.com/
>> 
>> 


Mime
View raw message