I tried hadoop-1.0.3 as well as 1.0.4. Both under pseudo cluster mode.
My understanding is that the previous.checkpoint directory should be created under secondary name node designated by "fs.checkpoint.dir"? On all occasions I am finding it under the namenode directory designated by "dfs.name.dir". It this something to do with Pseudo mode or my understanding is wrong? Can someone please help!
Below is my hdfs-site.xml file.
<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="configuration.xsl"?>
<!-- Put site-specific property overrides in this file. -->
<configuration>
<property>
<name>dfs.replication</name>
<value>1</value>
</property>
<property>
<name>dfs.name.dir</name>
<value>/home/hadoop/lab/hdfs/namenode</value>
</property>
<property>
<name>dfs.data.dir</name>
<value>/home/hadoop/lab/hdfs/datanode</value>
</property>
<property>
<name>fs.checkpoint.dir</name>
<value>/home/hadoop/lab/hdfs/secnamenode</value>
</property>
</configuration>
The below is high level directory structure for the daemons
hadoop@ubuntu:~/lab/hdfs$ ls -l
total 12
drwxr-xr-x 6 hadoop hadoop 4096 Mar 14 03:41 datanode
drwxrwxr-x 5 hadoop hadoop 4096 Mar 14 03:41 namenode
drwxrwxr-x 4 hadoop hadoop 4096 Mar 14 04:46 secnamenode
Below is the NameNode directory details
hadoop@ubuntu:~/lab/hdfs$ ls -l namenode
total 12
drwxrwxr-x 2 hadoop hadoop 4096 Mar 14 04:46 current
drwxrwxr-x 2 hadoop hadoop 4096 Mar 14 03:13 image
-rw-rw-r-- 1 hadoop hadoop 0 Mar 14 03:41 in_use.lock
drwxrwxr-x 2 hadoop hadoop 4096 Mar 14 03:34 previous.checkpoint
Below is the SNN directory details
hadoop@ubuntu:~/lab/hdfs$ ls -l secnamenode
total 8
drwxrwxr-x 2 hadoop hadoop 4096 Mar 14 04:46 current
drwxrwxr-x 2 hadoop hadoop 4096 Mar 14 03:46 image
-rw-rw-r-- 1 hadoop hadoop 0 Mar 14 03:41 in_use.lock
If you need any further details please let me know.
Thanks Rags