Fixed - re-ran merge again and all is well - a trip through the indexes
with Luke showed the original segments still there and the re-run took
care of it....
rp
RP wrote:
Upgrade proceeding on 0.9x - was able to parse and index just fine
after first halt, but now I get an error in the logs and an empty
query response in the browser. Only clue is in tomcat log which is
tossing up this nice error for me and since I'm an admin and not a
serious coder, the src shouts nothing obvious to me when I look at
those lines...:
From Tomcat logs:
Dec 16, 2006 4:23:36 PM org.apache.catalina.core.StandardWrapperValve
invoke
SEVERE: Servlet.service() for servlet jsp threw exception
java.lang.NullPointerException
at
org.apache.nutch.searcher.FetchedSegments.getSummary(FetchedSegments.java:159)
at
org.apache.nutch.searcher.FetchedSegments$SummaryThread.run(FetchedSegments.java:177)
Catalina.out give this weird result - in asterisks below:
2006-12-16 16:16:55,215 INFO NutchBean - query request from 192.168.2.2
2006-12-16 16:16:55,216 INFO NutchBean - query: hoho
2006-12-16 16:16:55,216 INFO NutchBean - lang: en
2006-12-16 16:16:55,220 INFO NutchBean - searching for 20 raw hits
***2006-12-16 16:16:55,372 INFO NutchBean - re-searching for 40 raw
hits, query: sex -site:"www.cooking.com" -site:"www.foodtv.com"***
2006-12-16 16:16:55,732 INFO NutchBean - found 102857 raw hits
How did those sites get in the query..?? I didn't have them there as
I only searched for hoho..??
Again - I'm behind a slow pipe so to re-crawl all of this converted
data is not an option in the short term.... Only change from default
8 to 9 procedure was the addition of the mapred.speculation=false
which cured the parse error and allowed me to continue and index.
Only other thing may be the linkdb not being put where it was told,
but I moved it over (it built fine) and the index operation ran
without issue.
--
rp
