Hmm.. I just removed the “crawl” directory (output directory) from the command and it works! I’m storing the output in a Cassandra cluster using Gora anyway. So I don’t think I want to store that on HDFS :) -- Manikandan Saravanan Architect - Technology TheSocialPeople
On 4 January 2014 at 11:06:56 pm, Ted Yu (yuzhih...@gmail.com) wrote: Can you pastebin the stack trace involving the NPE ? Thanks On Jan 4, 2014, at 9:25 AM, Manikandan Saravanan <manikan...@thesocialpeople.net> wrote: Hi, I’m trying to run Nutch 2.2.1 on a Haddop 2-node cluster. My hadoop cluster is running fine and I’ve successfully added the input and output directory on to HDFS. But when I run $HADOOP_HOME/bin/hadoop jar /nutch/apache-nutch-2.2.1.job org.apache.nutch.crawl.Crawler urls -dir crawl -depth 3 -topN 5 I’m getting something like: INFO input.FileInputFormat: Total input paths to process : 0 Which, I understand, is meaning that Hadoop cannot locate the input files. The job ends for obvious reasons citing the null pointer exception. Can someone help me out? -- Manikandan Saravanan Architect - Technology TheSocialPeople