A few of us have been running it nearly daily to test 0.90, but not in a way that's comparable to previous tests (those published in the YCSB paper being the only known public numbers) and using "weird" settings in order, for example, to make regions split all the time and see how we can make it fail :)
So to run it on 0.90 in a meaningful way, one has to run it on either 0.20 or 0.89 too and using the same hardware. I don't expect major differences going grom 0.89 to 0.90 tho, expect for operation that aren't benchmarked like cold startup/shutdown. J-D On Fri, Dec 24, 2010 at 5:19 PM, Otis Gospodnetic <[email protected]> wrote: > Hi, > > With HBase 0.90 nearly out, I'm wondering if anyone has run YCSB recently, so > we > can see how 0.90 compares to previous versions (not to mention other DBs)? > > Thanks, > Otis > ---- > Sematext :: http://sematext.com/ :: Solr - Lucene - Nutch - Hadoop - HBase > Hadoop ecosystem search :: http://search-hadoop.com/ > >
