Recovering local articles that are not visible

[Russ Allbery]

Nick Giles <nick.giles at ontology.com> writes:

> I'm trying to recover some articles which have been lost by our
> installation of inn. Our server is an internal only setup, so just a few
> local groups. We had a power cut a couple of days ago, and now one of
> the groups is reporting 'Bad article number' for all its (old)
> articles. I've had a poke around, and the articles are in the 'articles'
> folder area, and there are entries in the overview .DAT file that
> correspond. The 'active' file had its maxindex reset to 001, when it
> should have been 009 (yeah, not a huge group...)

> My feeling is that the articles are just missing from some index, but
> now I've run into a bit of a brick wall of understanding. Which of the
> various databases / indexes do I need to update, and how do I do that?
> Is it the .IDX file for the group, and is there a way to rebuild the
> content of that from the content of the 'articles' directory?

This is exactly the problem that tdx-util -R is designed to solve.  For
the groups affected, you want to follow the recipe in the tdx-util man
page under EXAMPLES:

       Rebuild the overview information for example.test from a
       traditional spool directory:

           tdx-util -R <patharticles in inn.conf>/example/test -n example.test

       The last command may be useful for recovering from a bad crash or
       corrupted overview information for a particular group, if you are
       also using the tradspool article storage method.

You should then run ctlinnd renumber <group> to update your active file
based on the newly rebuilt overview information.  You should be able to do
this while the server is running.

-- 
Russ Allbery (rra at stanford.edu)             <http://www.eyrie.org/~eagle/>

    Please send questions to the list rather than mailing me directly.
     <http://www.eyrie.org/~eagle/faqs/questions.html> explains why.