My this helps:
http://wiki.media-style.com/display/nutchDocu/setup+multiple+search
+sever
However this uses local file systems.
From my point of view it makes no sense to use search server and
ndfs since you need to stream data to the search server from a
external datanode to later on stream data to the tomcat server again.
In general ndfs indexes are slow as I'm typing.
Anyway check that your search server is setuped to use ndfs and the
path is a *absolute* path in the ndfs.
If you run search server and tomcat on one box (what makes no sense,
only for testing) also verify that your search-server.txt is located
in the ndfs, since I guess tomcat and search server share it
configuration files.
HTH
Stefan
Am 28.12.2005 um 01:32 schrieb Gal Nitzan:
Do I need to run server if I want to use the search to use NDFS?
Any way, in the nutch-site.xml which reside under tomcat the
serch.dir =
crawl and the name of the ndfs root is the same.
However, I still get 0 results though I know for sure there are
documents in the index.
On Wed, 2005-12-28 at 01:42 +0200, Gal Nitzan wrote:
Hi,
I have tried all available samples but was unsuccessful.
I am using the following command to start the server:
bin/nutch-daemon.sh start server 9003 crawl
I have setup a directory /hosts with the file search-servers.txt
which
contains localhost 9003
but the tomcat client does not connect to my search server at all.
Any idea what am i doing wrong?
Gal
---------------------------------------------------------------
company: http://www.media-style.com
forum: http://www.text-mining.org
blog: http://www.find23.net
