slave file is used by hadoop scripts to execute datanode on the slave
nodes. that is the only place, where you should specify new
datanodes. namenode currently does not know in advance about the
datanodes that are going to register.
On Dec 6, 2006, at 9:55 AM, John Wang wrote:
> To configure multiple data nodes, would I add to hadoop-
> default.xml or
> the slave file or both?