What is the Secondary NameNode?

A.
An alternate data channel for clients to reach HDFS, should the NameNode become too busy.

B.
A process that performs a checkpoint operation on the files produced by the NameNode.

C.
A data channel between the primary name node and the tertiary NameNode.

D.
A process purely intended to perform backups of the NameNode.

E.
A standby NameNode, for high availability.

Explanation:
http://wiki.apache.org/hadoop/FAQ#What_is_the_purpose_of_the_secondary_namenode.3F (3.2)