I'm running Hadoop 0.20.1+133 (Cloudera distro)
I tried setting up a multi-node Hadoop cluster and on executing the command:
hadoop jar /usr/lib/hadoop/hadoop-0.20.1+133-examples.jar grep input output
'dfs[a-z.]+'
I get:
09/10/27 20:39:21 INFO mapred.FileInputFormat: Total input paths to process
I've seen those errors when I was playing the values in the
core-site.xml, dfs-site.xml and mapreduce-site.xml.
Might be worth comparing your values to mine discussed in the thread
http://www.mail-archive.com/common-user@hadoop.apache.org/msg02522.html
which also represent 8G DN machines
Cheers