El Miércoles, 26 de Marzo de 2008, escribió: > On Wed, Mar 26, 2008 at 08:19:04PM +0100, Raúl Sánchez Siles wrote: > > Well, you're right, the thing is how this situation arises. TBH I don't > > really know how to get into this, I've tried several things, truncating a > > file that I know that aptitude will download to 0 bytes and also to stop > > it in the middle of a download and then truncating the result e.g. to > > half its downloaded size. apt-cache understood the situation and marked > > the file as corrupted before starting over to download the file. > > To be honest, that sounds perfectly correct behaviour. Agreed, but if it were always like this I wouldn't have placed this report ;)
>
> > Maybe there's some issue
> > on the apt-cache integrity DB if such a thing exists.
>
> Yes is does. Are you using checksumming (checksum=1 in config file)?
I'm not using this.
>
> I am still trying to get a handle on what you problem is. Has this
> situation arisen for you? Exactly what happened? Is it reproducible? If
> it is, enabling debugging info (debug=1 in the config file) will put
> lots of useful info into /var/log/apt-cacher/error.log.
>
> Mark
As far as I can remember I run aptitude as usual and some files(packages)
were reluctant to be downloaded. They took so much time and aptitude stalls
the download till it finally fails, approximately 5 minutes later or so. I
don't know what lend to that situation, maybe it has anything to do that I
also update a chroot on this system using apt-cache as well. So my main
system and the chroot uses the same cache.
Once I noticed the problem I tried to remove the affected file from the
apt-cache cache, and it all started to work again. After that I found that
apt-cacher-cleanup.pl would do a better work.
Another thing that may be of interest is that I'm not always connected to
the internet, so maybe the apt-cacher-cleanup.pl is doing anything wrong, but
this is unsupported conclusion.
I still don't have a method to reproduce it, but if happens again, I'll take
care to enable the debug option and send the relevant information.
I know I haven't been very useful with this reply, but if this problem
happens again I'm sure I'll try to corner it.
Thanks so far,
--
Raúl Sánchez Siles
----->Proud Debian user<-----
Linux registered user #416098
signature.asc
Description: This is a digitally signed message part.
