[Mimedefang] Another Feature Request

[Anthony Slay]

On Thu, 9 May 2002, David F. Skoll wrote:

> On Thu, 9 May 2002 17:40:19 -0700 (PDT) Anthony Slay
> <anthony at ocga.ucsd.edu> wrote:
> 
> > I, unfortunately, had to revert back to pre-mimedefang because I received
> > some "451 4.7.1 Please try again later" messages that are kinda
> > unacceptable when it comes to business here (bummer).  This happened
> > about 3 or so times today for about a minute at a time.  Spawning too
> > many slaves just works the cpu even harder.
> 
> Even with the latest beta, which tries to pace the spawning of slaves?
>

I was using version 2.9 because it was my fisrt "compiled" version
(installing that 18GB HD took priority) and since beta's come out so
often, I decided to wait for the non betas.

> > Anyway, getting to he point, it would be nice if Mimedefang could be run
> > on a different machine on the network so that machine "A" wanting to use
> > mimedefang's filtering ability just has to correctly rebuild their
> > sendmail configuration to use machine "B"'s running mimedefang processes.
> 
> In theory, it can be; you can supply a tcp connection for mimedefang's
> "-p" flag.  Milter has some syntax for doing this:
> 
>         mimedefang -p inet:port at hostname
> 
> So if you want mimedefang to listen on port 9989, I think you do this:
> 
>         mimedefang -p inet:9989
> 
> In the Sendmail config file, the connection is
> 
>         Xmimedefang, S=inet:9989 at 192.168.3.12, ...
> 
> assuming Mimedefang is running on 192.168.3.12
> 
> mimedefang and mimedefang-multiplexor and all the slaves must run on
> the same machine, but they can run on a different machine than sendmail.
> 
> --
> David.
>

Cool.  I'll try this.  Hopfully the new version and different server will
solve my delay and reject problems all at the same time.  Thanks for the
help!