amrad email problems

Robert E. Seastrom rs at seastrom.com
Sun Oct 14 22:54:25 CDT 2007


"Howard Cunningham" <howardc at macrollc.com> writes:

> Hi 
>
> At Tacos today there was a discussion about why some emails
> are not delivered.  I looked at the email that Mike posted
> tonight and found that there are multiple problems with the
> zone file setting for the amrad.org domain and email..
>
> Emails:
> The mx record specifies 204.29.138.195
> Emails are coming from: received: from atanasoff.rf.org
> ([204.29.138.196])
> That mismatch will cause more and more mail servers to not
> accept and/or reject emails.

Not well-configured nameservers unless there's SPF stuff that says
only receive from mx ("v=spf1 mx -all").  If you are aware of any mail
servers that come configured out of the box to refuse everything that
doesn't come from the proper mx, I'd appreciate a pointer.

> DNS Issue #1:
> Antanasoff.amrad.org is as an amrad name server but it is
> not listed with the registrar

This shouldn't cause a problem with resolving amrad.org (if the
reverse were true, i.e. the authoritative servers were a subset rather
than a superset of the registrar's records) it might cause slowness.

It is an inconsistency that ought to be fixed.  It is probably not the
source of mail slowness.

> DNS Issue #2:
> Antanasoff.amrad.org will do recursive lookups.  I recommend
> that recursion be disabled
>
> 204.29.138.196 reports that it will do recursive lookups

This presents a DNS cache poisoning issue.  It shouldn't be getting in
the way of resolution.  Agree with Howard that there should be no
recursive nameservers should also be authoritative nameservers, but
cavalierly disabling recursion on antanasoff without ascertaining that
nothing is depending on it for resolution might result in the IT
equivalent of having golf cleat marks on your tender bits.

So, I suspect that log analysis will help you divine what's unhappy,
if anything.

                                        ---Rob



More information about the Tacos mailing list