1 HBase uptime issues
Region sever fails due to overload, then GC pauses, then Zookeeper declared the region servers dead. Sometimes, the region server is out of memory
2 How many memory is needed?
data node + task tracker running on the same node, with 4G memory, is that enough?
3 How much data can be handled by the certain cluster? What is the ideal configuration and load for a region server
3TB data will be demonstrated in the way: 231 regions, 2055 stores, 1180 storeFiles adn 15MB storeFileIndex