Exploring HDFS Federation and Its High Availability - Cloudera Administration - page 122

Configuring HDFS Federation

For HDFS Federation to work, the datanodes need a way to identify the different namen-
odes in the cluster. There are several configuration parameters that need to be added to
hdfs-site.xml , which are listed in the following table:

Daemon Configuration Properties

dfs.nameservices

dfs.namenode.rpc-address

dfs.namenode.servicerpc-address

dfs.namenode.http-address

dfs.namenode.https-address

Namenode

dfs.namenode.keytab.file

dfs.namenode.name.dir

dfs.namenode.edits.dir

dfs.namenode.checkpoint.dir

dfs.namenode.checkpoint.edits.dir

dfs.namenode.secondary.http-address

Secondary namenode

dfs.secondary.namenode.keytab.file

We saw a few of these configuration properties in Chapter 2 , HDFS and MapReduce . The
following list defines a few of the important properties used to configure HDFS:

• dfs.nameservices : This property defines all the nameservices in the cluster.
The nameservice is a name used to uniquely identify an HDFS instance. The val-
ues are set as a comma separated list.
• dfs.namenode.rpc-address : This property defines the RPC address that
handles all client requests.
• dfs.namenode.https-address : This property defines the secure HTTP
server address for the namenode.
• dfs.namenode.keytab.file : This property defines the location of the
keytab file. A keytab file is a file that contains principals and encrypted keys. This
file is part of the Kerberos configuration, which we will explore in detail a bit later.