Re: [Exim] Hostname used in HELO

Pàgina inicial
Delete this message
Reply to this message
Autor: Exim Users Mailing List
Data:  
A: exim-users
Assumpte: Re: [Exim] Hostname used in HELO
[ On Monday, September 16, 2002 at 15:11:37 (+0100), Matt Saunders wrote: ]
> Subject: Re: [Exim] Hostname used in HELO
>
> Well, that's OK for me since I have the option of using TLSed ASMTP
> to a server on a fixed IP. But without that option I'd be stuck
> as I'd have to use a different relay host depending where I was
> plugged in (Easynet's at work, NTL's at home, and $RANDOMISP's when
> I roam elsewhere). This is why I prefer to be a "server" and send
> mail directly.


Well, that might be what you prefer to do, but that might not always be
what's allowed (and I don't just mean by policies local to the site
you're connected to, but also by policies for the remote mailer you're
attempting to deliver to).

It's a failing of your mail reader (and perhaps other tools) that you're
not able to conveniently re-configure them to match your current network
environment.

If your mail reader can only send via a local command-line tool then
whatever tool you use should be configurable to deliver to a specific
gateway host for further routing and next hop delivery. With Exim this
is done with a router which explicitly routes all non-local mail to a
specific gateway host. You might consider having your DHCP client, PPP
up scripts, etc., reconfigure this "smart host" router to properly match
your current network environment (eg. use the local SMTP gateway).

> Besides, I think it's inelegant to have a HELO/rDNS mismatch whether
> you're using a smarthost or not, so, given that it's still on the
> wishlist I'll be knocking something up to make use of Patrice's
> suggestion in the meantime. :-)


Well, that's a different question/issue, and not quite the right stating
of it either. The rule is that the HELO name "MUST" resolve to an A RR
that gives the source address the client-SMTP is connecting from.
Alternatly a domain literal can be given in the form of the IP address
for the source.

Whether or not the reverse rDNS must _also_ match, and whether or not it
even exists, is a somewhat separate question. General practice is that
IFF the rDNS exists then it "must" match. This is enforced by TCP
Wrappers and similar checks done for other services (eg. rsh, smail).

--
                                Greg A. Woods


+1 416 218-0098;            <g.a.woods@???>;           <woods@???>
Planix, Inc. <woods@???>; VE3TCP; Secrets of the Weird <woods@???>