Bill Cole <[email protected]> said: > > returning a null MX (MX 0 ".") with unbound locally if the upstream > server is responding NXDOMAIN on a MX query. > > This will cause loss of mail. >
I hear this from time to time, and I wonder how many people have actually tested or quantified this "loss of mail". Eighteen years ago I worked for a small ISP (200k customers) and we started blocking incoming email from domains that lacked an MX record. After 4 years we'd had exactly zero reports about it; none from the thousand of our users who hosted their own mail servers and used ours for relaying; and none from non-customers who could read my direct-dial phone number in the SMTP rejection message. (I did get occasional calls, but never about this.) This measure significantly reduced spam, so we left it in place, and extended it to reject outbound relaying to domains that lacked an MX. In the years since I quit working for that ISP, I've maintained my own mail server and had exactly one complaint about rejected email. That person had already fixed their issue by inserting an MX record in their DNS zone, and they just felt obligated to complain that I was not complying with the RFC. It is (and always will be) entirely valid to receive email at a FQDN which > has no MX record but does have a valid A or AAAA record. Today this would appear to be useful only for the miniscule proportion of domains that are "receive only", since a significant portion of outgoing mail from MX-less domains has been blocked for decades, or for delivery within closed networks where you can write your own rules. I guess support for MX-less domains will likely remain in the successor RFCs, because there will always be holdouts who will resist its removal, but from a global perspective, it was only ever necessary as a transition measure after the introduction of MX records, and could have been formally deprecated decades ago. -Martin PS: It's not exactly hard to add this to one's DNS: @ 999999999 MX 0 @ -- ## subscription configuration (requires account): ## https://lists.exim.org/mailman3/postorius/lists/exim-users.lists.exim.org/ ## unsubscribe (doesn't require an account): ## [email protected] ## Exim details at http://www.exim.org/ ## Please use the Wiki with this list - http://wiki.exim.org/
