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

Switch to Threaded View
Kafka >> mail # user >> Kafka + Maven


Copy link to this message
-
Re: Kafka + Maven
Thanks for the feedback everyone.

I do want to mention that the lack of action on this is not because we
don't care. Most all the committers have been almost entirely heads down on
getting replication, the network re-architecture, the log reimplementation,
and the system test framework done for 0.8 for the last 9 months or so.
That represents a pretty huge chunk of code (from around 13k loc to almost
19k loc). There just hasn't been a lot of slack time to spend on other
important things like the build system, website, documentation, etc. The
good news is that we are feature complete on 0.8 and though there is a lot
of hardening left to do, it does appear to work as we hoped, which is
incredibly exciting to me personally.

A second practical problem seems to be that almost none of the committers
seems to know either SBT or maven well enough to just sit down and do this.
So thanks a lot to people who have been trying to help us out.

-Jay
On Tue, Nov 20, 2012 at 3:27 PM, Chris Riccomini <[EMAIL PROTECTED]>wrote:

> Hey Guys,
>
> I was talking with Jay, and he recommended I forward some feedback along.
>
> I have been playing with Kafka 0.8 this week, and am feeling the pain in
> the lack of Maven support for it. Specifically, it'd be nice if this stuff
> were:
>
>   1.  In Apache's SNAPSHOT repository
>   2.  In some release repository.
>
> I am pretty sure that you guys are aware that it'd be nice to have, but I
> think I'd like to elevate the priority to blocker, or at least pretty
> critical.
>
> From a community adoption perspective, it's really damaging to the project
> to not have it easily accessible in a repository. To be blunt, if I hadn't
> heard of Kafka (and didn't know you guys), and I came across the page, and
> discovered that it had no maven coordinates, I'd move on to other solutions
> without a second glance. To me, it looks as bad as having no commits for
> the project in months. It looks unmaintained.
>
> If you look at what's already going on with the community, you'll see a
> lot of pretty rough things being done to try and get around this issue.
>
>   1.  Storm has its own publication: https://clojars.org/storm/storm-kafka
>   2.  SenseiDB has it committed in code, and runs a manual install. This
> makes it pretty rough if anyone DEPENDS on SenseiDB, as they, in turn, must
> do the same thing.
> https://github.com/senseidb/sensei/blob/master/sensei-gateways/pom.xml
>   3.  Druid is hosting its own:
> https://metamx.artifactoryonline.com/metamx/pub-libs-releases-local/kafka/core-kafka/
>   4.  It's been requested for more than a year:
> https://issues.apache.org/jira/browse/KAFKA-133
>
> This is only going to get worse when 0.8 is released, as it introduces
> even more unmanaged libraries (metrics, zkclient). I assume, currently,
> that releasing 0.8 is not contingent on getting those libraries published,
> but I think it should be. This is worth doing correctly. I don't think
> it'll be fun, but I really do think it's a big deal.
>
> Any chance we could get this on the road map soon? I think it falls under
> 2013 community adoption.
>
> Thanks!
> Chris
>