Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Hadoop >> mail # general >> [VOTE] - Establish YARN as a sub-project of Apache Hadoop

Copy link to this message
Re: [VOTE] - Establish YARN as a sub-project of Apache Hadoop

On Aug 13, 2012, at 2:01 PM, Tom White wrote:
> The committer list needs to be set up
> by editing https://svn.apache.org/repos/infra/infrastructure/trunk/subversion/authorization/asf-authorization-template.

Good point, thanks for pointing it out.

> It should have the initial contents of the MR committer list. When we split the committer lists for Hadoop into Common, HDFS, MR we made
> them all start out with the Hadoop list, so we should do the same for splitting YARN out of MR.
For a while now, we maintain distinct committer lists for HDFS & MapReduce and nominations for new committers are based on their contributions to the individual sub-project.

Following that I would like to propose the following.

Given that:
# YARN is a substantial (~100K LoC) and new body of work which has been around for nearly 18 months now.
# It is also very distinct from Hadoop MapReduce itself, as discussed in this thread, and wasn't derived from any of the existing Hadoop sub-projects.
I propose we define the list as the committers who have been active and have made significant contributions to the project.

A quick review of subversion leads me to suggest the following list (in lexical order).
If there is anyone who has contributed significantly to YARN and is not on that list, please accept my apologies in advance and let me know, I'll add you in.

* Alejandro Abdelnur
* Arun C. Murthy
* Chris Douglas
* Hitesh Shah
* Jonathan Eagles
* Luke Lu
* Mahadev Konar
* Robert Evans
* Siddharth Seth
* Sharad Agarwal
* Thomas Graves
* Vinod Kumar V

This approach seems consistent with what we have done historically with other new sub-projects such as Hive, Pig, HBase, ZooKeeper, etc. and with the Apache principle of Meritocracy.
Also, since we're starting with folks from 6 companies we already meet the Incubator criteria for diversity, which seems healthy.  

Obviously, additional contributors will be encouraged to join the sub-project via the normal mechanisms.  

Last, not least, most people on the MapReduce committer list are already part of the Hadoop PMC and hence they have svn rights if necessary - this doesn't change that and we continue to trust them.