dslreports logo
 
    All Forums Hot Topics Gallery
spc
Search similar:


uniqs
787
Floid2
join:2002-02-11
Ridgefield, CT

Floid2

Member

[OS X] OS X Mail vs. AT&T - SMTP issues?

Apparently this is a relatively 'random' SNAFU only biting small subsets of AT&T DSL customers when it does [or else everyone with a Mac has been unable to raise a stink due to lack of email?], but I'm short on time, clue, and access to the affected machine (or a Mac in general to poke at).

Does anyone happen to have any firm technical detail on what's causing the situation described in:

»SMTP auth issues for Torrington @sbcglobal.net user?

(Apparently rejecting username/password via SMTP when configured 'correctly,' webmail unaffected.)

»discussions.apple.com/th ··· /4388720 is linked therein regarding the same issue but is mostly speculation and chicken-waving.

Any assistance appreciated - apparently this has been going on for a couple weeks, but I'm loathe to put anyone on the phone to support on T's side.

mjf
MVM,
join:2000-08-05
New Orleans, LA

1 recommendation

mjf

MVM,

This thread solved my problem:

»BS Yahoo Mail and Mac Mail 6.x
kitsune
join:2001-11-26
Sacramento, CA

kitsune to Floid2

Member

to Floid2
The issue is that Mail.app is not getting a response from the SMTP server when you are using the default DNS servers. I have had success with both Google DNS and OpenDNS. I have also heard of ATT providing alternate servers with success but I'm not sure if they were their own or just Google or OpenDNS again.
Floid2
join:2002-02-11
Ridgefield, CT

Floid2

Member

Bit of a crosspost for others following along here - if you see my posts today in the original thread:

»SMTP auth issues for Torrington @sbcglobal.net user?

At first it appeared that if DNS is involved, it's seems more likely to be that nearly every home gateway on the planet won't resolve the CNAME for Mail.app... but then at least one Apple forum poster suggested it's a quietly 'known issue' - and having just read up on the infinite variations of SMTP authentication within a 'single' method, I can buy that, too.

(Or at least, I'm taking the fact that they apparently pointed the poster to an unauthenticated server to be significant.)