On 2/26/21 2:10 PM, borg@uu3.net wrote:
Hmm right... Somehow I tought that having that special Null MX will silently discard message... I dont know why...
So, RFC 7505 is pretty much even pointless in my opinion. You have to do more.. to pretty much achieve the same.. Its just easier to not having MX on subdomains that does not serve as email destinations.. Less records in DNS..
It should mean that there is no attempt to deliver email, even if the domain has an A or AAAA record.
---------- Original message ----------
From: Grant Taylor via NANOG <nanog@nanog.org> To: nanog@nanog.org Subject: Re: Newbie Question: Is anyone actually using the Null MX (RFC 7505)? Date: Fri, 26 Feb 2021 12:03:37 -0700
On 2/26/21 11:46 AM, borg@uu3.net wrote:
Well, I bet my legacy system will bounce it for example... What specifically is the bounce?
I thought the purpose of the Null MX was to do two things:
1) Provide as an MX that can't be connected to. 2) Serve as a signal to things that know how to interpret it that no mail is to be expected.
I would expect that some server, if not the MSA, /would/ generate a bounce /because/ the email to the domain is undeliverables.
I cant speak about Sendmail, qmail, Exim.. when they started supporting it. My Sendmail boxes have been dealing with the Null MX just fine. The aforementioned bounce is /expected/ to tell the sender that the destination address is bad.
So, In my opinion changing already working standards in a way that they arent full compat with old systems is imo bad aproach. IMHO there is little, if any, effective difference between the Null MX and an MX pointing to an unresolvable name or an non-routed IP. They cause a hard / fast failure in an early upstream MTA thus induce a bounce.
Depending on the MSA, the delivery problem may even be presented to the user as they are submitting the message to the MSA.
-- Grant. . . . unix || die