>
> I wanted to contribute a while back, but there was hardly any support
> on new contributors and especially the ones who wanted to write larger
> features and bugs.
>


> Upstream pre-built packages for production use (Start rhel7, maybe ubuntu)
>

A similar, developer-focused idea I've been toying with is hosting a Docker
image that contains a pre-built thirdparty directory. Depending on the
machine, running thirdparty/build-if-necessary.sh could take on the order
of hours, after which the build itself might slap on some more.

Republishing a thirdparty Docker image whenever we update thirdparty could
help lower the barrier to entry for a developer wanting to write and test
more meaningful patches. The idea then would that we could expose some
build-with-docker.sh, or have users develop, build, and test, all in the
Docker image itself. Would be interested in hearing if there's interest in
that.

On Mon, Jul 23, 2018 at 4:19 PM Mike Percy <[email protected]> wrote:

> Thank you for the feedback!
>
> I think that pointing to the newbie label <https://s.apache.org/OcI6> on
> JIRA (not always tagged reliably) and the design docs
> <https://github.com/apache/kudu/tree/master/docs/design-docs> section of
> our Git repo would be a good start for improving CONTRIBUTING.adoc or
> http://kudu.apache.org/docs/contributing.html (rather than talking about
> style guides and stuff).
>
> Mike
>
> On Mon, Jul 23, 2018 at 2:36 PM Atri Sharma <[email protected]> wrote:
>
> > A potential ideas list, with internals architecture would be great.
> > On Mon, Jul 23, 2018 at 11:40 PM Mike Percy <[email protected]> wrote:
> > >
> > > Hi Atri, do you have suggestions on what we could do better next time?
> > >
> > > Mike
> > >
> > > On Mon, Jul 23, 2018 at 10:56 AM Atri Sharma <[email protected]> wrote:
> > >
> > > > I wanted to contribute a while back, but there was hardly any support
> > > > on new contributors and especially the ones who wanted to write
> larger
> > > > features and bugs.
> > > >
> > > > Just my 2c
> > > > On Mon, Jul 23, 2018 at 11:16 PM Sailesh Mukil
> > > > <[email protected]> wrote:
> > > > >
> > > > > On Tue, Jul 17, 2018 at 7:37 PM, Mike Percy <[email protected]>
> > wrote:
> > > > >
> > > > > > On Tue, Jul 17, 2018 at 12:22 PM Grant Henke
> > > > <[email protected]>
> > > > > > wrote:
> > > > > >
> > > > > > > I have started a document for blog post ideas/topics here:
> > > > > > >
> > > > > > >
> > https://docs.google.com/document/d/12QFRIhNDMoOI1kOQBgch64xYJ9t6U
> > > > > > byVt1D3NaTl7lI/edit?usp=sharing
> > > > > > >
> > > > > >
> > > > > > Nice list, Grant. Actually I think that quarterly email would
> > probably
> > > > make
> > > > > > for a better blog post instead and I've added it as a suggestion
> on
> > > > that
> > > > > > doc.
> > > > > >
> > > > > > On Tue, Jul 17, 2018 at 12:04 PM Mauricio Aristizabal <
> > > > [email protected]
> > > > > > >
> > > > > > wrote:
> > > > > >
> > > > > > > I was disappointed that Strata SJ 2018 didn't have a single
> > session
> > > > on
> > > > > > > Kudu, there were no committers in attendance that I could tell,
> > and
> > > > it
> > > > > > > wasn't being highlighted at all in the Cloudera booth.  Between
> > > > Strata
> > > > > > and
> > > > > > > ScalaDays I must have enthusiastically mentioned the product to
> > 15
> > > > people
> > > > > > > and none had heard of it.
> > > > > > >
> > > > > >
> > > > > > Hmm, that is disappointing, and a bit surprising. Perhaps
> everybody
> > > > thought
> > > > > > everybody else was going to submit... actually I had intended to
> > > > submit a
> > > > > > talk proposal to Strata this year but got busy and missed the
> > > > deadline. :(
> > > > > >
> > > > > > I wonder if folks using Kudu would like to present on their use
> > case?
> > > > I'm
> > > > > > sure conference-goers would like to hear from more people using
> > Kudu
> > > > "in
> > > > > > anger" (hopefully not angrily).
> > > > > >
> > > > > > On Tue, Jul 17, 2018 at 2:59 PM Sailesh Mukil
> > > > <[email protected]
> > > > > > >
> > > > > > wrote:
> > > > > >
> > > > > > > A suggestion to add on to the easily downloadable pre-built
> > > > packages, is
> > > > > > to
> > > > > > > have easily accessible/downloadable example test-data that's
> > fairly
> > > > > > > representative of real world datasets (but it doesn't have to
> be
> > too
> > > > > > > large). Additionally, we can write tutorials in kudu/examples/
> > that
> > > > use
> > > > > > > this test data, to give new users a better feel for the system.
> > > > > >
> > > > > >
> > > > > > That sounds useful. Any ideas on where we could find such a data
> > set?
> > > > > >
> > > > >
> > > > > Starting with a small scale factor of TPC-H and TPC-DS might not be
> > a bad
> > > > > idea.
> > > > >
> > > > >
> > > > > > On Tue, Jul 17, 2018 at 11:59 AM Tim Robertson <
> > > > [email protected]>
> > > > > > wrote:
> > > > > >
> > > > > > > ++1 on the mini cluster
> > > > > > > Perhaps include a docker image build at the same time which
> > > > presumably
> > > > > > > wouldn't be much effort?
> > > > > > >
> > > > > >
> > > > > > I'm not really sure there is a lot of overlap between creating a
> > Docker
> > > > > > image and the kind of relocatable artifacts I'm trying to build,
> > aside
> > > > from
> > > > > > the actual compiling part. But I think it would be valuable for
> > Docker
> > > > > > users to be able to easily pull down a Kudu image.
> > > > > >
> > > > > >
> > > > > > > l'll be happy to contribute on the Java / maven related parts
> to
> > > > that. I
> > > > > > > will use this for the testing framework for the Apache Beam
> > KuduIO
> > > > and
> > > > > > will
> > > > > > > certainly help test / write a blog.
> > > > > > >
> > > > > >
> > > > > > I don't really know how to handle the Maven part where we unpack
> > the
> > > > > > tarball and set it up somewhere so we can invoke it from the
> > > > > > KuduMiniCluster. Maybe it that would require writing a custom
> Maven
> > > > plugin?
> > > > > >
> > > > > > I'd love to see a blog post about how to use Kudu with Beam!
> > > > > >
> > > > > > Mike
> > > > > >
> > > >
> >
>


-- 
Andrew Wong

Reply via email to