[leafnode-list] Re: Fetchnews process takes very long

[Peter J Ross]

On Saturday 18 October 2008 10:31:51 Torsten Bronger wrote:
> Hallöchen!
>
> I run fetchnews as a cron job every 10 minutes.  Every other hour or
> so, it takes *much* longer than usual (15 minutes instead of approx
> 40 seconds).  I suspect that it does some sort of database cleanup
> at this time.

The "database cleanup" is performed by texpire (man 8 texpire). The frequency 
is controlled by cron.

> One reason is probably that I've subscribed to many groups over the
> last months but unsubscribed from most of them.  However, fetchnews
> still gets news from them -- which is not totally bad, because
> sometimes I re-visit such groups.
>
> But can one move this *very* I/O-intensive cleanup of fetchnews to
> the night hours?

Yes, investigate the cron settings for the "news" user and the system (man 5 
crontab).

> Or, can I tell fetchnews which groups I really 
> won't want to read again?

If you delete groups from /var/spool/news/interesting.groups/ (or wherever 
the "interesting.groups" directory appears in your installation), leafnode 
will stop fetching messages.

-- 
PJR :-)