🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: demon.ip.developers
Article #1049

Re: KA9Q and SMTP timing

#1049
From: Mark Rogers
Date: Sat, 13 Apr 1996 00:00
64 lines
3719 bytes
In article <829315028snz@siberia.demon.co.uk>
   Philip McNab <philip@siberia.demon.co.uk> writes:
> 
> Alternatively, one method that would be guaranteed to work is to
> rewrite your .wrk files so that the first line always reads
> 'post.demon.co.uk'.  You could do this by editing them manually
> (I do it with a lot of my mail), or by using a simple program
> to do it: if you're really keen, and the 'smtp wait 0' option
> doesn't work, I could run something up quite quickly for you to
> place in your demon.bat.

Preferably, such a utility would not rewrite mail which is going to 
*.demon.co.uk; ideally it would be configurable to allow decisions based 
on who you are sending to and, even, when you are connecting. This was 
all discussed a while back, as I had grand plans for such a project 
myself, but time seems to have vanished recently. I've still got all the 
discussion here if anyone wants it, though.

> remote sites has gone more quickly.  It seems to me that Demon could
> do with more mail-servers (both sales and support staff have
> acknowledged this in phone conversations), but that's another issue:

The mail-servers, AIUI, are not properly scalable yet, so this isn't 
trivial. (The problem is that Demon's customers run a variety of 
software, much of which cannot cope with a large number of simultaneous 
connections from all the servers. I'm not sure how such a problem could 
be resolved - no doubt Demon are working on it. ISTR that having 4 punts 
broke some software.)

> A further complication is that Demon's mailservers sometimes build
> up an accumulated backlog of outgoing messages: at times in the last
> few months this has reached 24 hours, so while your mail may get out
> of your site quickly, it may end up sitting in a traffic jam at
> Finchley when other roads are clear.  

Mail sent via post will generally be slower than if sent direct; it's 
really just a question of by how much. In most cases the delay is trivial 
(esp. when compared with the reduction in online time). In some cases 
this delay can indeed get quite large; this hasn't happend for some time, 
though.

One thought that just occured to me; presumably if all mail is sent via 
post then KA9Q's "smtp batch" command can be used safely?

In passing, recent advice was to use relay-1.mail.demon.net 
(158.152.1.140) as a smart host if post.demon.co.uk was unavailable. 
Sending all mail via post and then having "smtp gateway" also configured 
to post is a bit pointless, so I now use relay-1 as my "smtp gateway" 
instead. YMMV.

> Maybe someone will correct me, but even if there were some way of 
> recording the time data you seek, it would not be meaningful at low 
> orders of magnitude.

It would probably also not be valid at the same time next week, or a 
different time on the same day, or...

I would go for "all non-Demon mail via post" as a starting point; perhaps 
tweaking this as time goes on.

Mark.
-- 
int MeaningOfLife (void) { return 42; }   // mcr@holly.demon.co.uk
function MeaningOfLife :integer; begin MeaningOfLife := 6 * 9 end;
Technical Details
Message-ID: <829429175snx@holly.demon.co.uk>
Path: arc1-ca.pugleaf.net!unknown!not-for-mail
References: <829306771snz@bhhpharm.demon.co.uk> <829315028snz@siberia.demon.co.uk>

Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 10bb05,c5341e4805b90b63
X-Google-Attributes: gid10bb05,public
From: Mark Rogers <mcr@holly.demon.co.uk>
Subject: Re: KA9Q and SMTP timing
Date: 1996/04/13
Message-ID: <829429175snx@holly.demon.co.uk>
X-Deja-AN: 147462713
x-mail2news-path: disperse.demon.co.uk!post.demon.co.uk!holly.demon.co.uk
x-nntp-posting-host: holly.demon.co.uk
references: <829306771snz@bhhpharm.demon.co.uk> <829315028snz@siberia.demon.co.uk>
organization: Unn Ltd.
newsgroups: demon.ip.developers
x-mailer: cppnews $Revision: 1.43 $