Hi all, 1) I wanted to know how strong the coupling between HDFS and MapReduce (programming abstraction) in Hadoop is. Can someone throw some light on the protocols used between HDFS and JobTracker/TaskTracker/Namenode interactions. Any pointer on this would be of great help!
2) Does the Hadoop system utilize the local storage directly for any purpose (without going through the HDFS) in clustered mode? Thanks, Matthew John
