[Discuss] Strange sendmail (and postfix) spam issue: accepting fail "from" myself?
Derek Atkins
warlord at MIT.EDU
Thu Apr 14 11:25:59 EDT 2016
Hi,
Derek Martin <invalid at pizzashack.org> writes:
> On Tue, Mar 29, 2016 at 02:09:15PM -0400, Derek Atkins wrote:
>> So now I want to focus on Sendmail. Any sendmail guru's out there?
>
> Hello Derek!
>
> Did you ever find a solution for this for Sendmail? I would never
> call myself a Sendmail guru, especially now--but I recall I spent some
> time on trying to solve this a while back and failed. It seems the
> Sendmail folks believe that fixing this is a bad idea, because it can
> block legitimate mail e.g. if someone at your site sends mail to
> someone at another site that has a .forward file that points to an
> address at your site. In case that's not clear:
>
> From: foo at example.com
> To: bar at example.org
>
> And bar at example.org has a .forward file that forwards to
> baz at example.com.
>
> Apparently, this message will get lost. This seems like it should be
> a fixable problem, but pffft.
>
> I did just find this recipe, which appears to be outdated:
>
> http://www.sendmail.org/~ca/email/examples/Ted.html
>
> I also thought SPF and/or DMARC would fix this, but I never got around
> to trying to set any of that up...
Alas, no, I never did figure this out. :(
In my case, I know that the "forward back to myself" is never going to
happen. This is just a mailman server, so all mail is either
originating locally or being relayed through mailman. There should
never be a remote connection where MAIL FROM is my domain.
Of course it doesn't differentiate between a connection from 127.0.0.1
and a connection from elsewhere. :( So it's blocking mailman too when
I put those blocks in.
I think I might just switch to postfix when I have time.
-derek
--
Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
Member, MIT Student Information Processing Board (SIPB)
URL: http://web.mit.edu/warlord/ PP-ASEL-IA N1NWH
warlord at MIT.EDU PGP key available
More information about the Discuss
mailing list