hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Erik Bengtson" <e...@jpox.org>
Subject RE: DataNucleus HBase plugin
Date Wed, 13 May 2009 23:49:17 GMT

Google wrote and maintains the DataNucleus plugin. We provide them support
for their development.

Back to HBase, I gave up with Windows and I have now a Linux one running.
Indeed, I will take a look at MiniHBaseCluster.

And the good news is that the HBase plugin for JDO/JPA/REST is ready and a
nightly build will be available to download in 1 or 2 days from now.

Besides the windows stuff, everything else worked well. Congratulations.

Erik Bengtson

-----Message d'origine-----
De : saint.ack@gmail.com [mailto:saint.ack@gmail.com] De la part de stack
Envoyé : mardi 12 mai 2009 23:43
À : hbase-dev@hadoop.apache.org
Objet : Re: DataNucleus HBase plugin

On Tue, May 12, 2009 at 1:01 AM, Erik Bengtson <erik@jpox.org> wrote:

> I'm a DataNucleus developer (implementation of JDO/JPA/REST APIs).
> Currently, we
> support Google BigTable, and I would like to extend our support for HBase.

Out of interest, how do you add support to a proprietary product?

I've been trying to run HBase 0.19 (using cygwin), but when I run the
> MyClient
> API example, it cannot find anything on port 60000.

For sure its up?  It says so in hbase logs?

HBase in cygwin purportedly works (one of our devs runs its this way).

> I would also like to setup test cases for our plugin, so it would be
> interesting
> if we run hadoop+hbase in embedded mode.

See the unit tests.  See MiniHBaseCluster in particular.  This runs a JVM
hosting an hbase master and N regionservers.  In trunk, we'll also start a
separate JVM to host zookeeper; maybe we should try running this inside the
same minihbase jvm?

If you can't figure it, come back here so we can help you out.


View raw message