Thanks for the tip. Actually I realized that the code was working properly but it take some time until the INDEX folder get visible on hadoop.
Does anyone know ways to improve the indexing speed by using map-reduce ? addIndex from IndexedTableAdmin is quite expensive for large tables... Thanks, Adrian ________________________________ From: Jean-Daniel Cryans <[email protected]> To: [email protected] Sent: Tue, January 26, 2010 7:38:14 PM Subject: Re: index creation problem in a fully distributed hbase Adrian, If it doesn't throw any stack trace, you could try running jstack'ing the process and see where the thread is stuck. J-D On Tue, Jan 26, 2010 at 9:58 AM, Adrian Popescu <[email protected]> wrote: > Hello All, > > I try to create some indexes in hbase with a code tested on a non-distributed > installation. The problem is that my application which builds the indexes > hangs immediately after creating partially the index: the .regioninfo file is > created but not the actual index (__INDEX__ is not populated at all). I guess > that there is a setting that I miss somewhere on the fully distributed > installation but I don't know which. On the other hand I can load, scan > tables from the hbase, so hbase is working... > > Thank you, > Adrian > > >
