[quote user="Peter Strömblad"]I've checked dessergroup2.demon.co.uk and it's not a valid domain[/quote]

It's a sub-domain.

[quote]The domain demon.co.uk also yields a number of warnings[/quote]

They look like minor issues even with your tools.  Certainly nothing that Mercury should stumble over.

I'm getting an odd error  when sending mail to one specific domain.

The domain is <dessergroup2.demon.co.uk>

The connection history shows:

---------------------------------------------------------------------------

14:20:13: processing job MO000249
14:20:13: processing job MO0002C6
14:21:17: processing job MO000242
Resolved CN for 'dessergroup2.demon.co.uk' to 80.177.153.214
Connecting to 80.177.153.214
Resolved CN for 'dessergroup2.demon.co.uk' to 80.177.153.214
Connecting to 80.177.153.214
Connection error.

Resolved CN for 'dessergroup2.demon.co.uk' to 80.177.153.214
Connecting to 80.177.153.214
Connection error.

Connection error.
---------------------------------------------------------------------------

I do not understand what is going on. 

The quoted IP is the IP of that subdomain and not that of the MX record(s) that do exist.

Why is  this happening?

A quick glance without my tools yields that the domain is missing its SOA record, but if that should do this ...

Looks all right to me:

[code]paul@briod:~$ host dessergroup2.demon.co.uk
dessergroup2.demon.co.uk has address 80.177.153.214
dessergroup2.demon.co.uk mail is handled by 10 punt-1.mail.demon.net.
dessergroup2.demon.co.uk mail is handled by 10 punt-2.mail.demon.net.
paul@briod:~$ host  punt-1.mail.demon.net
punt-1.mail.demon.net has address 194.217.242.248
punt-1.mail.demon.net has address 194.217.242.75
paul@briod:~$ host  punt-2.mail.demon.net
punt-2.mail.demon.net has address 194.217.242.248
punt-2.mail.demon.net has address 194.217.242.75[/code]

As you say, there's no reason why MercuryE is trying to deliver to the A address unless it's getting a bad response from the name servers.

If you have MercuryE configured to use the Windows DNS, try changing it to specific good servers.

I think I have got to the bottom of this - though what is actually happening I do not understand.  Looks like a bug in the code resulting from invalid data.

I noticed in the MercuryE log messages like "501 <@dessergroup2.demon.co.uk>: no local part"

This took me to the Aliases which I had set up earlier where I found that an alias had nothing in front of the @

Fixing that seems to have solved the problem.

Hmm, if this is the case, then MercuryE should return a failure ... but ... I don't recall how MercuryE interacts with the alias file...

I'll ask David to have a look at this.

sorry, can't let this go quite - I've checked dessergroup2.demon.co.uk and it's not a valid domain. What you have is a stub zone with an A address and two mx records. A zone like that is what you'd use internally only. The domain demon.co.uk also yields a number of warnings. I've used the following tools

http://www.dnsstuff.com/

http://dnscheck.iis.se/?lang=en