I have three question: 1 How large a region will be, if i want to use map-reduce to analyze the data in hbase table. 2 How many region in a region server is suitable? 3 If the rowkey in my application is not dispersive such as md5 or hash, how can i control the hotspot of region split?
thx
