On Thu, 12 Sep 1996, Ilya Ketris wrote:
> Wouldn't it be nice to restrict incoming SMTP
> connections on the basis of load average, not just
> maximum connection count? Number of simultaneous
> connections often says little about the general load.
Suggestion noted.
On Fri, 13 Sep 1996, Piete Brooks wrote:
> The minimal one (just restrict to N coons) is a 20 line perl script,
> but I am using a sexy one which records the caller info as well,
> instances of calls being rejected, and tailoring based on calling host
> (e.g. I allow 20 calls from within the dept [our MUA doesn't cope well if
> mail isn't accepted] and 5 from outside)
>
> Anyone else interested ?
Piete, I'm prepared to distribute this with Exim if you think that would
be useful.
On Fri, 13 Sep 1996, Nigel Metheringham wrote:
> However this is building everything into exim. Maybe some form of
> general parameter gathering interface could be added...? This might
> enable the whole system portability aspect of load average and disk
> space etc to be side stepped - maybe in a similar way to INN
I think Piete's approach may be the way forward here - don't build
everything into Exim, but have a separate "gateway" that can do any
checks it likes before passing the call on to Exim.
On Fri, 13 Sep 1996, Ilya Ketris wrote:
> While the whole list looks undoubtedly reasonable, I would give the
> parameters already present in other mailers special treatment,
> because people would expect them to be present.
That is certainly a valid point. That is why I noted the suggestion
about load average, and have added it to the Wish List.
--
Philip Hazel University Computing Service,
ph10@??? New Museums Site, Cambridge CB2 3QG,
P.Hazel@??? England. Phone: +44 1223 334714