Problem opening checksum file

[Stefan Groschupf <sg...@media-style.com>]

Hi,
what is meaning this and how to fix this ? :-o
051015 221418 Problem opening checksum file: java.io.IOException:  
Cannot find filename /user/myuser/db/current/part-00012/.index.crc.  
Ignoring.
Looks like it isn't critical at all but I was wondering why this can  
happen.

Thanks for any hints.
Stefan

Re: Problem opening checksum file

[Doug Cutting <cu...@nutch.org>]

Stefan Groschupf wrote:
> what is meaning this and how to fix this ? :-o
> 051015 221418 Problem opening checksum file: java.io.IOException:  
> Cannot find filename /user/myuser/db/current/part-00012/.index.crc.  
> Ignoring.
> Looks like it isn't critical at all but I was wondering why this can  
> happen.

Files that were created before Nutch stored checksums (last Thursday) 
will not have checksums and will trigger this warning on open.

This warning is also sometimes triggered for Lucene indexes, even newly 
created ones.  Here the problem is that Lucene indexes are not written 
with NutchFileSystem, and thus may not get checksummed, but they are 
read with NutchFileSystem, which desires checksums.  We are not yet able 
to directly write Lucene indexes with the NutchFileSystem API, since, in 
one place, Lucene overwrites data in a file, and the NutchFileSystem API 
does not permit that.

Doug