[
https://issues.apache.org/jira/browse/NUTCH-2553?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16435844#comment-16435844
]
Sebastian Nagel commented on NUTCH-2553:
----------------------------------------
Hi [~omkar20895], the URLs used are invalid (at least, outside Java: Firefox
fails to open http://nutch.apache.org:-1/), so we have to change this
definitely. There was a reason, indeed, why fit.url wasn't used here: it
partially masked a bug in QueueFeeder
([c93d908:L87|https://github.com/apache/nutch/commit/c93d908bb635d3c5b59f8c8a22e0584ebf588794#diff-e2801ecd867ab2df559ae0c91c138cb1L87])
which just added the same URL/CrawlDatum pair again and again to the queue(s).
See [Hadoop Object Reuse
Pitfall|https://cornercases.wordpress.com/2011/08/18/hadoop-object-reuse-pitfall-all-my-reducer-values-are-the-same/]
for an explanation. Anyway, good that the awkward URLs did uncover it: it also
affected the CrawlDatum values which would have been harder to detect. Fix/PR
is in preparation.
> Fetcher not to modify URLs to be fetched
> ----------------------------------------
>
> Key: NUTCH-2553
> URL: https://issues.apache.org/jira/browse/NUTCH-2553
> Project: Nutch
> Issue Type: Bug
> Components: fetcher
> Affects Versions: 1.15
> Reporter: Sebastian Nagel
> Priority: Critical
> Fix For: 1.15
>
>
> Fetcher modifies the URLs being fetched (introduced with NUTCH-2375 in
> [c93d908|https://github.com/apache/nutch/commit/c93d908bb635d3c5b59f8c8a22e0584ebf588794#diff-847479d08597eb30da1c715310438685R253]:
> {noformat}
> FetcherThread 22 fetching http://nutch.apache.org:-1/ (queue crawl
> delay=5000ms)
> {noformat}
> which makes it hard to trace the URLs in the log files and likely causes
> other issues because URLs in CrawlDb and segments do not match
> (http://nutch.apache.org/ in CrawlDb and http://nutch.apache.org:-1/ in
> segment).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
