Just sharing my experience with setting the search directory for the nutch webapp. This is a leading cause of the disappointing "Hits 0-0 (out of about 0 total matching pages)" message.
I had a situation like Noah Silverman: > On Thu, 2009-12-17 at 16:32 -0800, Noah Silverman wrote: > >> Hello, >> >> Just to summarize. >> >> 1) Nutch crawl completes without error. >> >> 2) I can search from command line and see results. (Assume this means >> that index is created.) >> bin/nutch org.apache.nutch.searcher.NutchBean foobar >> >> 3) Tomcat configured through nutch-site file to point to nutch/crawl >> directory >> >> 4) catalina.out logfile indicates that tomcat is opening nutch/crawl >> 2009-12-16 22:00:39,740 INFO SearchBean - opening indexes in >> /home/noah/Documents/nutch/crawl/indexes >> >> 5) No results when searching in web front end >> >> 6) No errors in the logs >> >> Is there some way to debug this? Perhaps more verbose logging? >> >> Thanks!!! >> >> -N The log message in 4 is only somewhat helpful since if anything goes wrong, nothing will be said. Noah's problem was that he needed to point to the top level directory. My case was that I needed to set the permissions correctly. I had crawled as root so the crawl directory was root:root with permissions 544. (at least readable) I moved it to $TOMCAT/work and gave it ownership $TOMCAT_USER:$TOMCAT_GROUP with permissions 755. Now it works. In any case, the nutch web app will simply log at info that it's opening indexes at $DIR. If permissions are wrong, or the directory doesn't exist, it will say nothing, not even at debug logging. No exceptions will be thrown.
