hadoop-yarn-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Alejandro Abdelnur <t...@cloudera.com>
Subject READY: Apache Jenkins job to create Hadoop 2.x release artifacts
Date Fri, 31 Jan 2014 19:22:27 GMT

The Apache Jenkins job to build release artifacts for Hadoop 2 is ready (in
its first incarnation). The job URL is:


I hope this will make moot all concerns about user accounts and about the
hardware ownership being used to generate Hadoop 2 release artifacts.

*Andrew & Arun,*

Because the two of you are driving/coordinating the release of Hadoop
2.3.0, you'll be the first ones to use the release script and job. If you
run into any issues or you have ideas on how to improve it, please don't
hesitate to let me know; I'll be more than happy to help.

One thing I ask you both is to update the "How to Release" wiki page to
reflect all necessary current steps to build a release, specially the
manual steps.

Having this up to date and complete will help others to take on the Release
Manager tasks without being suck into a time sink.


If Bigtop wants to consume release artifacts from this job on regular
basis, we'll need to modify the job to run periodically. Fell free to do
so, just make sure that when running because of a cron trigger the RC_LABEL
used states so, i.e.: NIGHTLY<YYMMDD>. And all the produced artifacts will
be labeled with it.

Also, if for Bigtop is more convenient to have artifacts without SNAPSHOT
JARs even if the  Hadoop POMs still have SNAPSHOT in the branch, I have an
idea on how we could do that (doing a build with an additional patch -being
fetched- on top of the branch).

Thanks and cheers to all.


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