In recent versions of Mac OS X, a default Hadoop configuration such as from Homebrew<https://github.com/Homebrew/homebrew/blob/a49c2fa1244032f12ba0e1121a934bcecfe10182/Library/Formula/hadoop.rb>raises errors on some operations:
$ hadoop version Hadoop 1.2.1 Subversion https://svn.apache.org/repos/asf/hadoop/common/branches/branch-1.2 -r 1503152 Compiled by mattf on Mon Jul 22 15:23:09 PDT 2013 >From source with checksum 6923c86528809c4e7e6f493b6b413a9a This command was run using /usr/local/Cellar/hadoop/1.2.1/libexec/hadoop-core-1.2.1.jar When querying HDFS, Hadoop reports a realm error: $ hadoop fsck / -files -bytes 2014-03-12 10:55:48.330 java[12749:1703] Unable to load realm info from SCDynamicStore ... This happens because realms are not configured by default. Setting HADOOP_OPTS in hadoop-env.sh fixes the realm error: export HADOOP_OPTS="-Djava.security.krb5.realm=-Djava.security.krb5.kdc=" $ hadoop fsck / -files -bytes ... Another error occurs when attempting to start a namenode: $ hadoop namenode 14/03/12 11:25:25 ERROR namenode.NameNode: java.lang.IllegalArgumentException: Does not contain a valid host:port authority: file:/// This is because the core-site.xml bundled with Hadoop fails to specify an address. Adding an address property fixes this: <property> <name>fs.default.name</name> <value>hdfs://localhost:8020</value> </property> $ hadoop namenode ... 14/03/12 11:27:53 INFO ipc.Server: IPC Server listener on 8020: starting These manual steps to correct the configuration might scare off newbies. In the future, could Hadoop come with better defaults, for a better out of the box experience for newbies? -- Cheers, Andrew Pennebaker apenneba...@42six.com