I am having a hard time getting nutch 1.3 to run in a pseudo distributed mode on Windows Server 2008 sp2.
I spent a week messing with hadoop version 0.20.203.0 and I have come to the conclusion that it is not possible to start the task tracker due to an issue in RawLocalFileSystem.java(515). It is trying to do a setuid by setting the permissions to 0700. It is not possible to do this in this version of Windows. I think it has to do something with the least privileged user model for permissions. To bypass this I went back to version 0.20.2 and I was at least able to start all the components and run the examples hadoop jar and get it to finish. Fast foward to running nutch. I am able to get it to start an inject, but it fails with some long garbage looking path while trying to write to log files. What should the paths for the directories in the hadoop conf files be? /cygdrive/r/hadoop or r:/cygdrive/r/hadoop or file:///r:/cygdrive/r/hadoop The only thing on the lists I could find talking about this were from this message thread: http://www.mail-archive.com/nutch-user@lucene.apache.org/msg14205.html -- View this message in context: http://lucene.472066.n3.nabble.com/Nutch-1-3-Cygwin-paths-tp3336911p3336911.html Sent from the Nutch - User mailing list archive at Nabble.com.
