RE: [Exim] Rejecting due to reverse DNS lookup failture

Αρχική Σελίδα
Delete this message
Reply to this message
Συντάκτης: Lewis Ashton
Ημερομηνία:  
Προς: Terry Edhouse, exim-users
Αντικείμενο: RE: [Exim] Rejecting due to reverse DNS lookup failture
Hi,

sorry to drag up an old one again, but I'm getting behind on reading the
list :-(

Whilst I worked for an ISP we had the same issue. I implemented the DNS
lookups during an upgrade. Marketing went nuts about this because we got so
many complaints. The standard arguments failed to cut it. When we did some
analysis of the logs, we found that 1 in 5 unique IP addresses were failing
the lookups. So we had to abandon the check.

As a minor aside, when people did what I asked and contacted their ISP, I
spend a lot of time doing support for ISPs out there to fix their DNS.

The whole thing raised some serious questions in my mind about how far you
should go to satisfy customers when non-RFC compliance is the issue. I know
where I stand, but commerce doesn't really match that in any shape or form.

Lewis


...Lewis wanders off into the distance ranting about disclaimers, spam and
some of the even more dubious things companies want mail systems to do...

> -----Original Message-----
> From: Terry Edhouse [mailto:tje1@ohm.york.ac.uk]
> Sent: 27 June 2000 14:25
> To: exim-users@???
> Subject: [Exim] Rejecting due to reverse DNS lookup failture
>
>
>     At the moment we refuse to accept mail from hosts where 
> a reverse DNS
> lookup fails to supply a hostname for the calling IP address, 
> or if the
> returned name gives a different IP address when it is 
> looked-up (or the
> lookup fails).

<snip>

> 2000-06-27 14:06:53 connection from [209.38.156.44] refused (failed to
> find host name from IP address)


<snip>

>     My question is - how common is this practice amongst 
> members of the
> list, and how do you deal with the users who give you grief 
> because mail
> to them is not getting through?

>
>
> --Terry Edhouse
> Department of Electronics
> University of York

<snip>