Hello everyone, I have a cluster from 8 datanodes and a namenode. When I start teragen program everything works OK, the data is generated. But when I start terasort program, seems like that only 2 datanodes do the job. And everything is so slow. I've tried with only 10 records and cluster generated sort in few seconds, but with bigger number, it stucks around 15%.
Do you have any idea why this is so? I'm using Hadoop 0.20.2 and Ubuntu 8.10. Thank you.
