Phil Pennock <exim-users@???> said, in message
20070603212600.GB33153@???:
> On 2007-05-31 at 08:49 +0100, Alun wrote:
> > On 22nd May we upgraded our mail servers to exim 4.67. Since then
> > we've been having problems with mail from wales.ac.uk. Mail from
> > them originates from 194.80.131.5 which has multiple reverse DNS
> > records.
>
> I can see one which might be relevant. Exim 4.64:
> PH/48 When a host name was looked up from an IP address, and the
> subsequent forward lookup of the name timed out, the host name was
> left in $sender_host_name, contrary to the specification.
Thanks! I looked at the changelog, but didn't think hard enough. The
lookup doesn't timeout and, in fact, when I do a lookup using "host",
"nslookup" and "dig" it works. So I discounted this as a likely reason.
Exim is obviously using something rather more low-level than "host"
and choking on the wildcard.
> Which for me yields the below, during setup. Given that this is Exim
> complaining about no IP found, I suspect that the DNS changed around
> this time too and it's not really related to the Exim change.
> Unfortunately the administrators of 131.80.194.in-addr.arpa. are not
> using date-encoding for the SOA serial number, so I can't confirm
> this.
The administrator insists that there was no change and I'm inclined to
believe him. At 08:22 on 22nd May we were receiving mail with no
problems. By 08:50 we were having problems. The date stamp
on /usr/sbin/exim is 08:33.
I'll get back to him and suggest that his DNS is, at best, wildly
non-standard and, at worst, illegal. I've already told him he should
expect this to become an increasing problem if it's not an exim bug.
Thanks everyone,
Alun.
--
Alun Jones auj@???
Systems Support, (01970) 62 2494
Information Services,
University of Wales, Aberystwyth
