You have to think if this is a realistic load scenario, will you really have 200 threads executing as fast as possible per hbase instance? Remember you realistically need a 10 node cluster.
Even with a high writer count will you really be writing back to back? Also think about your mutation semantics, with the history and multi column features hbase gives you, are you sure you'll need to do what you're doing? Hbase is not sql, and take advantage of the features we provide. On May 9, 2009 4:48 PM, "Guilherme Germoglio" <[email protected]> wrote: But with 200 threads, the problem is back. =\ On Sat, May 9, 2009 at 8:44 PM, Guilherme Germoglio <[email protected] >wrote: > maybe you've just spotted the problem! I've raised > hbase.regionserver.handler.count to 100 and ... -- Guilherme msn: [email protected] homepage: http://germoglio.googlepages.com
