[jira] [Commented] (NUTCH-2975) Generate 0 partition when used with sitemap

["Sebastian Nagel (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/NUTCH-2975?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17637810#comment-17637810 ] 

Sebastian Nagel commented on NUTCH-2975:
----------------------------------------

The Generator log message only says that no records were selected. Could you
- provide the full logs, i.e. the Hadoop task logs
- look into the CrawlDb after SitemapProcessor was run and verify that it's definitely of size 0
- share the sitemap

... or at least some snippets. Otherwise it's hard to figure out the reason.

Just one assumption which could be the reason: SitemapProcessor writes the modification time, priority and fetch interval from the sitemap to the CrawlDb. This is only possible if you can trust these values. A modification time in the future, or a score of 0.0 (not larger than the default of the property generate.min.score) could cause that the URLs from the sitemap are not eligible for crawling.

> Generate 0 partition when used with sitemap
> -------------------------------------------
>
>                 Key: NUTCH-2975
>                 URL: https://issues.apache.org/jira/browse/NUTCH-2975
>             Project: Nutch
>          Issue Type: Bug
>          Components: generator, sitemap
>    Affects Versions: 1.16
>         Environment: Hadoop 3.1.1
>            Reporter: Lucas Pauchard
>            Priority: Major
>
> *Issue*
> We are facing strange issue since we have updated our Proxmox from 7.2-4 to 7.2-11 which host the VMs/containers used for our Hadoop cluster.
> When we are using the sitemap component to add URLs, the generator process doesn't work. It generates 0 partition.
> But if we call a second time the generator process, this time the generator actually create a partition segment.
> It happens only when we use the sitemap process. If we use only the Injector process, this issue doesn't happen.
> I checked the logs and the generator just seems to find no record in the crawldb. It is like the crawldb wasn't available or the files are locked.
> *Here is the command used :*
> Sitemap :
> {code}
> hadoop jar <job> org.apache.nutch.crawl.Generator crawl_000_111/crawldb crawl_000_111/segment -numFetchers 2 -mapper=1 -reducer=1 -noFilter -noNorm -force
> {code}
> It returns as expected :
> {panel:title=Sitemap output}
> 2022-11-23 10:37:22,194 INFO util.SitemapProcessor: SitemapProcessor: Total records rejected by filters: 0
> 2022-11-23 10:37:22,195 INFO util.SitemapProcessor: SitemapProcessor: Total sitemaps from HostDb: 0
> 2022-11-23 10:37:22,195 INFO util.SitemapProcessor: SitemapProcessor: Total sitemaps from seed urls: 1
> 2022-11-23 10:37:22,196 INFO util.SitemapProcessor: SitemapProcessor: Total failed sitemap fetches: 0
> 2022-11-23 10:37:22,196 INFO util.SitemapProcessor: SitemapProcessor: Total new sitemap entries added: 151
> {panel}
> Generetor :
> {code}
> hadoop jar <job> org.apache.nutch.crawl.Generator crawl_000_111/crawldb crawl_000_111/segment -numFetchers 2 -mapper=1 -reducer=1 -noFilter -noNorm -force
> {code}
> 1st time it returns :
> {code}
> 2022-11-23 11:25:15,202 WARN crawl.Generator: Generator: 0 records selected for fetching, exiting ...
> {code}
> 2nd time it returns :
> {code}
> 2022-11-23 11:27:43,007 INFO crawl.Generator: Generator: Partitioning selected urls for politeness.
> 2022-11-23 11:27:44,009 INFO crawl.Generator: Generator: segment: crawl_000_111/segment/20221123112744
> ...
> 2022-11-23 11:28:34,061 INFO crawl.Generator: Generator: finished at 2022-11-23 11:28:34, elapsed: 00:01:53
> {code}



--
This message was sent by Atlassian Jira
(v8.20.10#820010)