Regarding checksum error in hadoop in my latest PR.

Previous Topic Next Topic
 
classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Regarding checksum error in hadoop in my latest PR.

Omkar Reddy
Hello dev@,

I am facing an EOFException in the file TestGenerator.java and I cannot get my hands on the way in which I can solve it. The Exception is as follows :

  1. 2017-08-09 12:57:06,026 WARN fs.FSInputChecker (ChecksumFileSystem.java:<init>(157)) - Problem opening checksum file: file:/tmp/hadoop-omreddy/mapred/temp/generate-temp-16af5bc1-1a80-412b-b0ca-481c82877f3b/fetchlist-0/part-r-00000. Ignoring exception:
  2. java.io.EOFException

I cannot understand the reason for it. This PR[1] is the part of an effort to upgrade Nutch to use new MapReduce API. 

Please find the detailed log of the test here[0]. Any suggestions/help would be appreciated. 

Thanks,
Omkar.

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Regarding checksum error in hadoop in my latest PR.

Sebastian Nagel
Hi Omkar,

the test fails because nothing is generated, i.e. the tests detected a broken functionality. :)

Best is to test generator from the command-line whether they behave as expected, e.g. (with a
non-empty CrawlDb):

  $ bin/nutch generate path/to/crawldb path/to/segments/
  ...
  Generator: segment: path/to/segments/20170811152636
  Generator: finished at 2017-08-11 15:26:37, elapsed: 00:00:03

  $ tree path/to/segments/
  path/to/segments/20170811152636
  `-- 20170811152636
      `-- crawl_generate

The folder crawl_generate is empty!  Generator is complex with multiple steps working in temporary
folders. Eventually, it's only the final copying which is broken. But I see no other way as to debug
the fetch list generation to find out what the reason is.

I strongly recommend to test also all other tools from command-line. For the bulk of them just run
a sample crawl via bin/crawl.

Best,
Sebastian


On 08/09/2017 11:56 AM, Omkar Reddy wrote:

> Hello dev@,
>
> I am facing an EOFException in the file TestGenerator.java and I cannot get my hands on the way in
> which I can solve it. The Exception is as follows :
>
>  1. 2017-08-09 12:57:06,026 WARN fs.FSInputChecker (ChecksumFileSystem.java:<init>(157)) - Problem
>     opening checksum file:
>     file:/tmp/hadoop-omreddy/mapred/temp/generate-temp-16af5bc1-1a80-412b-b0ca-481c82877f3b/fetchlist-0/part-r-00000.
>     Ignoring exception:
>  2. java.io.EOFException
>
>
> I cannot understand the reason for it. This PR[1] is the part of an effort to upgrade Nutch to use
> new MapReduce API.
>
> Please find the detailed log of the test here[0]. Any suggestions/help would be appreciated.
>
> Thanks,
> Omkar.
>
> [0] https://paste.apache.org/e1cQ
> [1] https://github.com/apache/nutch/pull/188

Loading...