huge news system architecture
Marek Iwanejko
M.Iwanejko at elka.pw.edu.pl
Mon Feb 25 12:06:59 UTC 2002
On Mon, 25 Feb 2002, Katsuhiro Kondou wrote:
> In article <Pine.SOL.4.30.0202242244400.13803-100000 at mion.elka.pw.edu.pl>,
> Marek Iwanejko <M.Iwanejko at elka.pw.edu.pl> wrote;
>
> } 1) 2-3 Heavy machines (let's say 4 CPU, 12GB RAM) as reader server(s)
> } with disk arrays. One server is master, rest are slave.
> } It's expensive, but full redundant, nice to maitenace.
>
> >From my experience, I guess, these should work for a few
> thousands clients simultaneously.
Thanks =)
> As to caching, what I'm worrying about is how the cache
> server know expired/cancelled articles. Some clients may
> not care, but others may complaint, in case that the article
> should be already gone.
Caching frontends are 'expired' by periodic (cron;) deleting oldest
files from cache spool. Ie. diablo cache server works that way.
The other problem is, that is similar to 'tradspool' with a lot of files
:-(
Regards,
Marek
--
_________________________________________________________________
Marek Iwanejko * M.Iwanejko at elka.pw.edu.pl *** in arp we trust **
More information about the inn-workers
mailing list