-All MAP Jobs(Java Custom Map Reduce Program) are assigned to one Node why?
I am using cdh4 with MRv1 . When I am running any hadoop mapreduce
program from java , all the map task is assigned to one node. It suppose
to distribute the map task among the cluster's nodes.
Note : 1) My jobtracker web-UI is showing 500 nodes
2) when it is comming to reducer , then it is sponning into
other node (other then map node)
Can nay one guide me why it is like so