>
> s2=`ls -d crawl/segments/2* | tail -1`

You cannot use "ls -d" to get the name of the generated segment directory
because the crawl directory exists on the HDFS and not your local
filesystem. Use "bin/hadoop dfs -ls crawl/segments |sort" instead.

Best,
Siddhartha Reddy

On Tue, Apr 22, 2008 at 6:06 AM, Jason Boss <[EMAIL PROTECTED]> wrote:

> >  Whole web?  What for, if not a secret?
>
> Tinkering and perhaps more.  I used nutch back in the day but dang you
> guys have come a long ways!
>
> >  Suggestion: don't run things as root.
>
> I know :)
>
> >  Have you formatted the filesystem?
>
> Yes, I formatted the file system as per a tutorial I found online:
> bin/hadoop namenode -format
>
> >  Can you run bin/hadoop fs -ls /user/root/crawl ?
> >
> [EMAIL PROTECTED] search]# bin/hadoop fs -ls /usr/root/crawl
> Found 0 items
>
> Doesn't look so good...
>
> >  Oh, if you have not injected any URLs, there is nothing to crawl in
> your crawldb.
> >  Run bin/nutch and you will see "inject" as one of the options.
> >
> bin/hadoop dfs -put urls urls
>
> I did a dfs -ls and it appears there.  For whole web indexing I was used
> to:
>
> bin/nutch generate crawl/crawldb crawl/segments -topN 1000
> s2=`ls -d crawl/segments/2* | tail -1`
> echo $s2
> bin/nutch fetch $s2
> bin/nutch updatedb crawl/crawldb $s2
>
> With hadoop what changes?  Do i just point to the virtual file system?
>
> Thanks a ton!
>
> Jason
>



-- 
http://sids.in
"If you are not having fun, you are not doing it right."

Reply via email to