Re: mail basics on a debian system
On Thu, Dec 12, 2002 at 01:26:57PM +0100, Robert Land wrote:
> On Wed, Dec 11, 2002 at 04:04:32PM -0500, Derrick 'dman' Hudson wrote:
> > Note that the data stream looks the same, except there is no response
> > from the server. The "batched" comes from the intent that the data
> > stream will be saved and then later fed into an MTA. An example
> > usage, as Colin indicated, is for a high-availability server to accept
> > the messages via SMTP (over TCP) and store the client machine's SMTP
> > commands in a file. The messages will be collected as they arrive,
> > then at some point the file will be transferred to a low-availability
> > server for delivery. The low-availability server receives all the
> > messages at once (as a "batch" job) via BSMTP.
>
> Derrick, please forgive me if this question sounds stupid ( I know
> I have to reread a lot of network stuff), but I have the impression
> that the high-availability server in Colins example acts somehow
> as a relay mashine with a buffer feature added to it?
> Does this mean the mashine of Colins friend does no verification,
> filtering or whatever, just collects the debian-list messages and sends
> the whole bunch on demand?
I was just coming up with a random example, no need to analyse it in
depth ...
But anyway, the high-availability server in question is effectively a
buffering relay, yes, if you accept having mail pulled rather than
pushing it out as a form of relaying. My friend's machine is completely
uninvolved in mail going through that server. The only filtering I do is
running procmail and spamassassin on my mail server at home, which is
exactly the same filtering I do on normal mail.
> What is the exact definition of a high-availability server?
One that you can deliver mail to reliably. A box on the wrong end of an
ADSL link without a static IP address doesn't fit that description.
Cheers,
--
Colin Watson [cjwatson@flatline.org.uk]
Reply to: