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

Switch to Threaded View
Bigtop, mail # user - BigTop : Build Targets


Copy link to this message
-
Re: BigTop : Build Targets
Sean Mackrory 2013-06-10, 21:54
> So I guess big top is a non standard project in that , although it builds
artifacts, those artifacts can already be built from other places.

In addition to what Roman said, the artifacts built by Bigtop are generally
better integrated with the target operating system. We provide additional
init scripts, friendlier wrappers in /usr/bin, integration with common
package managers, and we follow the conventions used in other Linux
software packages for file locations, etc.

But you're right that Bigtop is a non-standard project in that's (possibly)
the first Apache project that's more like a distribution of components than
one of the components themselves. In my opinion the value that Bigtop
provides by trying to unify the efforts of upstream projects exceeds the
value of the artifacts themselves.
On Mon, Jun 10, 2013 at 1:51 PM, Roman Shaposhnik <[EMAIL PROTECTED]> wrote:

> On Mon, Jun 10, 2013 at 1:40 PM, Jay Vyas <[EMAIL PROTECTED]> wrote:
> > So I guess big top is a non standard project in that , although it builds
> > artifacts, those artifacts can already be built from other places.
>
> The key word here is -- they can only be built. They are not available
> by default. Here's what I'm trying to say -- if today you download
> the default binary convenience artifact of Hadoop 1.x, HBase 0.94.x
> and Pig 0.11 you won't be able to even start the combination up.
> You have to rebuild the corresponding projects in order to make
> them work with each other.
>
> > However, in addition to building artifacts, it actually tests those
> artifacts between
> > one another - and the fact that it does this is actually one of the
> "major"
> > artifacts - because that process does the integration testsing for the
> > matrix of hadoop ecosystem tools.
>
> I'll stick with my Linux distro analogy. We're a Bigdata management
> distro. Our kernel is Hadoop and our userland is all the other projects.
> We produce the packages for end users to consume and we also
> maintain code that facilitates production of those packages (build,
> deploy, test).
>
> > Anyways... just thinkinng out loud while getting a feel for this awesome
> > peice of work.  Very useful and wish i knew of it earlier.  after a few
> > years of hadooping around and banging my head against integration
> ecosystems
> > various components.
> > Hope this helps others that are new to the big top project.!!
>
> We're definitely lacking good documentation if you think you can help --
> that
> will be VERY much apprecaited!
>
> Thanks,
> Roman.
>