How separate hadoop secondary namenode from primary namenode?

问题

all

I want to ask, now I'm running hadoop 2.6.0. So how can I separate this secondary namenode from the primary one? What's the configuration?

Have I use one additional computer to become a secondary namenode, or it can be in a datanode?

I need your suggest, thanks...

回答1:

NameNode, Secondary NameNode, DataNodes are just names given to "machines" based on the job they perform. In a "ideal" distributed enviornment, they all can and should reside in separate machines.

The only requirement for a secondary NameNode to be on another machine is : Secondary NameNode and NamdeNode should be connected so that The FSImane & EditLogs can be copied from NameNode to Secondary NameNode.

来源：https://stackoverflow.com/questions/28713804/how-separate-hadoop-secondary-namenode-from-primary-namenode