The following is a new meeting request:
Subject: Fedora "Big Data" Packaging and Guidelines & Pitfalls
Organizer: "Tim St Clair" <tstclair(a)redhat.com>
Time: Wednesday, October 2, 2013, 2:00:00 PM - 3:00:00 PM GMT -06:00 US/Canada Central
Invitees: matt(a)redhat.com; pmackinn(a)redhat.com; rrati(a)redhat.com; bigdata(a)lists.fedoraproject.org
As our SIG grows, we've had many questions about details around packaging.
As a result, we would like to openly invite folks to join in a conversation.
Conference Dial-in Number: (605) 475-4000
Host Access Code: 680076*
Participant Access Code: 680076#
Use the Freenode #fedora-bigdata channel for other-chatter.
Some links worthy of looking at prior to the conversation:
I'll try to update the content on the wiki below prior to our conversation:
I'm joining BigData SIG this week! I've been working with Hadoop on the
IBM Linux Technology Center since 2012. On the past months, I've been
building/testing/developing Sqoop, PIG, Zookeeper, HBase and Hadoop
on x86_64, IBM POWER and IBM SystemZ.
I'm looking forward to collaborating with you, expanding Fedora Hadoop's
ecosystem. Initially, I'll be primarily focused on porting packages to IBM
POWER using OpenJDK, but feel free to ping me anytime on freenode
(lrangel) to discuss ideas or suggestions.
IBM Linux Technology Center
I am Rodrigo Trujillo and work for IBM Linux Technology Center in
Brazil. I am looking forward to contribute with the packaging of some
Hadoop components for PPC64 and X86_64.
I have worked with build and fixes for Chukwa, Avro, Lucene and other
components... Also, I have some experience with RPM packaging, Koji,
mock, etc. from previous projects that I worked here, internally.
My name is Paulo Vital and I'm introducing myself to contribute in Big
Data SIG. I'm Software Engineer at IBM Linux Technology Center based in
Brazil and I've been working with Hadoop ecosystem since Oct,2012.
I have some experience with development of the Apache HBase, Ambari and
Zookeeper projects, including some patches submitted to community, and
the build process of Hadoop ecosystem components using Jenkins and
Apache Bigtop in our internal development environment.
So, I'd like to help Fedora community specially in HBase and Ambari (but open for other components) focused in Power (ppc64) and x86_64.
Best regards, Paulo (pvital@freenode).
Paulo Ricardo Paz Vital <pvital(a)linux.vnet.ibm.com>
IBM Linux Technology Center
In my investigation into packaging hbase, the upcoming release 0.96
really wants to build/work against hadoop 2.1.0. I imagine this is the
case for various parts of the hadoop ecosystem that are working on being
packaged for Fedora.
I'm going to start the process for updating the hadoop package to the
2.1.0-beta release. There will be 1 more spin of 2.0.5-alpha to clean
up some remaining issues and build compatibility with the new java
packaging tools and then I'll begin the 2.1.0-beta upgrade process
unless there are objections.
I've put together an HBbase packaging wiki and linked it off the main
Big Data SIG packaging page.
It seems light on missing dependencies, but my runs through the test
suite with mostly native packages haven't been encouraging. Hopefully
that will improve as the missing dependencies are packaged.