This happens to me if I don't do a clean build or there is a jar somewhere on the path containing an older hadoop version. Try "ant clean" before building and check for older releases. Please let me know if that helps. You may also run into a problem related to HADOOP-1212 after formatting the name-node.
Thanks, --Konstantin Zhu Huijun wrote:
Hi, I have a problem running Hadoop on our cluster. After I formated the namenode by executing "./bin/hadoop namenode -format", and started the dfs by executing "./bin/start-dfs.sh", I can access the webpage of namenode at http://localhost:50070, but the number of live data nodes is 0. If I run "./bin/stop-dfs.sh", there are messages of "no datanode to stop" after each ip addresses of the slave nodes listed in the /conf/slaves file. I found the following error in the log files: ERROR org.apache.hadoop.dfs.DataNode: org.apache.hadoop.dfs.IncorrectVersionException: Unexpected version of storage directory /tmp/hadoop-zhuhj/dfs/data. Reported: -11. Expecting = -10. However, I check the file of /tmp/hadoop-zhuhj/dfs/data/current/VERSION, it says "layoutVersion=-10". I appreciate that anyone can give me some hint about what the problem is. Thanks! Best wishes!
