Re: [spfbis] RFC7208 4.6.4 Interpretation - MX Lookup Count Inconsistencies

[Jan Schaumann <[e-mail]>]

Tim Wicinski <[e-mail]> wrote:

> I also feel that technology stacks have matured over time.  Jan, do you see
> real world examples
> of a domain with 11 MX servers?

I just happened to be looking at the MX records for
the top 1M domains for other reasons, and counting
domains that have > 10 MX records, I find at least 265
domains.

The winning price goes to the apparently rather aptly
named domain 'everymailbox.com', with 398 MX records.
Second prize goes to 'preciseify.com' (266 MX
records), 3rd to 'rm02.net' (235 MX records). :-)

Those are obviously rare outliers, of course.

But I'm actually less concerned about domains with >
10 MX records (since the RFC seems reasonably clear
here to immediately fail), and more worried about
domains that have close to 10 total lookups, but where
counting the lookups resulting from turning MX results
into IP addresses would bump them over the limit of
10.

As illustrated by Mark's initial mail (and my own
misunderstanding), there are implementations that will
lead to such domains' SPF records as being marked
invalid when they shouldn't be.

(The last time I looked at SPF records of popular
domains[1], I found >8K domains with >10 total
lookups.)

-Jan

[1] https://www.netmeister.org/blog/spf.html