1. 在masters文件中添加 Secondary节点的主机名。
*注：masters文件用于指定secondary的主机而不是namenode，slaves用于指定datanode和tasktracker，
namenode由core-site.xml fs.default.name指定，jobtracker由mapred-site.xml mapred.job.tracker指定
2. 修改hdfs-site.xml文件
<property>
<name>dfs.http.address</name>
<value>${your-namenode}:50070</value>
<description>Secondary get fsimage and edits via dfs.http.address</description>
</property>
<property>
<name>dfs.secondary.http.address</name>
<value>${your-secondarynamenode}:50090</value>
<description>NameNode get the newest fsimage via dfs.secondary.http.address</description>
</property>
*注：
实际上dfs.http.address只在secondary设置，dfs.secondary.http.address只在namenode上设置即可，为了便于管理，集群所有机器同样配置
采用默认端口(namenode:50070,secondary:50090)时可以省略该配置
2. 修改core-site.xml文件
<property>
<name>fs.checkpoint.period</name>
<value>3600</value>
<description>The number of seconds between two periodic checkpoints.</description>
</property>
<property>
<name>fs.checkpoint.size</name>
<value>67108864</value>
<description>The size of the current edit log (in bytes) that triggers a periodic checkpoint even if the fs.checkpoint.period hasn't expired. </description>
</property>
<property>
<name>fs.checkpoint.dir</name>
<value>${Hadoop.tmp.dir}/dfs/namesecondary</value>
<description>Determines where on the local filesystem the DFS secondary namenode should store the temporary images to merge.If this is a comma-delimited list of directories then the image is replicated in all of the directories for redundancy.</description>
</property>
*注:该配置在secondary设置即可，为了便于管理，集群所有机器同样配置
3. 重启hdfs，检查是否正常启动
(*注：这一步也可以不重启hdfs，在secondary上直接 sh $HADOOP_HOME/bin/hadoop-daemon.sh start secondarynamenode 启动secondaryNamenode)
(1)重启
sh $HADOOP_HOME/bin/stop-dfs.sh
sh $HADOOP_HOME/bin/start-dfs.sh
(2)检查uri
:50070/ #检查namenode
:50090/ #检查secondary
(3)检查目录
检查dfs.name.dir namenode:/data1/hadoop/name
current
image
previous.checkpoint
in_use.lock #主要看时候有这个文件，文件时间戳表示namenode启动时间
检查fs.checkpoint.dir secondary:${hadoop.tmp.dir}/dfs/namesecondary
current
image
in_use.lock #主要看时候有这个文件，文件时间戳表示secondnamenode启动时间
(4) 检查checkpoint是否正常
为便于测试，调整参数fs.checkpoint.period=60，fs.checkpoint.size=10240
对hdfs做一些文件增删操作，看${dfs.name.dir}/current/edits 和 ${fs.checkpoint.dir}/current/edits的变化