Aargh. The emr is executing but there is no feedback other than knowing 6 maps are executing 3 in running state and 3 in pending (for 5 mins now). No logs. This totally sucks. I am burning 6 c1 medium machines. I should be getting atleast 12 mappers running as c1s are 2 core.
I would rather go back to clouderas cluster with the hadoop tracker interface Robin
