Dennis Kubes wrote:
> I had a somewhat difficult time figuring out how to get hbase started.  
> In the end, it was pretty simple.  Here are the steps:
> 
>   1. Download hadoop from svn, untar to directory say ~/hadooptrunk and
>      compile through ant.
>   2. Move the build hadoop-xx directory to where you want to run it,
>      say ~/hadoop
>   3. Set the hadoop tmp directory in hadoop-site.xml  (as default all
>      other variables should be file)
>   4. Copy scripts from ~/hadooptrunk/src/contrib/hbase/bin to
>      ~/hadoop/src/contrib/hbase/bin
>   5. Format hadoop dfs through  ~/hadoop/bin/hadoop namenode -format
>   6. Start the dfs through ~/hadoop/bin/start-dfs.sh  (logs are
>      viewable in ~/hadoop/logs by default, don't need mapreduce for hbase)
>   7. Go to the hbase directory ~/hadoop/src/contrib/hbase
>   8. Hbase default values are fine for now, start hbase with
>      ~/hadoop/src/contrib/hbase/bin/start-hbase.sh (logs are viewable
>      in ~/hadoop/logs by default)
>   9. Enter hbase shell with ~/hadoop/src/contrib/hbase/bin/hbase shell
>  10. Have fun with Hbase
>  11. Stop the hbase servers with
>      ~/hadoop/src/contrib/hbase/bin/stop-hbase.sh.  Wait until the
>      servers are finished stopping.
>  12. Stop the hadoop dfs with ~/hadoop/bin/stop-dfs.sh
> 
> Hope this helps.

Did you try to run it with LocalFS / Cygwin, and if so, did you notice 
any peculiarities? I tried this once, and first the start-hbase.sh 
script wouldn't work (missing log files? it looked like some variables 
in paths were expanded in a wrong way), and then when I started the 
master and a regionserver by hand, it would complain about missing map 
files and all requests would time out ... I gave up after that and moved 
to HDFS.

-- 
Best regards,
Andrzej Bialecki     <><
  ___. ___ ___ ___ _ _   __________________________________
[__ || __|__/|__||\/|  Information Retrieval, Semantic Web
___|||__||  \|  ||  |  Embedded Unix, System Integration
http://www.sigram.com  Contact: info at sigram dot com