In the context of configuring a Hadoop cluster for HDFS High Availability (HA), ‘fencing’ refers to:

A.
Isolating a failed NameNode from write access to the fsimage and edits files so that is cannot
resume write operations if it recovers.

B.
Isolating the cluster’s master daemon to limit write access only to authorized clients.

C.
Isolating both HA NameNodes to prevent a client application from killing the NameNode
daemons.

D.
Isolating the standby NameNode from write access to the fsimage and edits files.

Explanation:
A fencing method is a method by which one node can forcibly prevent another node
from making continued progress.
This might be implemented by killing a process on the other node, by denying the other node’s
access to shared storage, or by accessing a PDU to cut the other node’s power.
Since these methods are often vendor- or device-specific, operators may implement this interface
in order to achieve fencing.
Fencing is configured by the operator as an ordered list of methods to attempt. Each method will
be tried in turn, and the next in the list will only be attempted if the previous one fails. See
NodeFencer for more information.
Note:
failover – initiate a failover between two NameNodes
This subcommand causes a failover from the first provided NameNode to the second. If the first
NameNode is in the Standby state, this command simply transitions the second to the Active state
without error. If the first NameNode is in the Active state, an attempt will be made to gracefully
transition it to the Standby state. If this fails, the fencing methods (as configured by
dfs.ha.fencing.methods) will be attempted in order until one of the methods succeeds. Only after
this process will the second NameNode be transitioned to the Active state. If no fencing method
succeeds, the second NameNode will not be transitioned to the Active state, and an error will be
returned.
org.apache.hadoop.ha, Interface FenceMethod

HDFS High Availability Administration, HA Administration using the haadmin command