At 12:41 PM 8/24/2004 -0400, Karl Prince
<spf(_dot_)pobox(_at_)princeweb(_dot_)com> wrote:
On Mon, 23 Aug 2004 13:00:01 -0600,
administrator(_at_)yellowhead(_dot_)com wrote:
The EHLO our
sendmailNT server uses by default is the physical server name <SERVER3>,
and is not configurable using the built in UI. It has passed
all SPF checks
that I have run against it.
For a little experiment, I suggest you try and add your
machines domain details to the hosts file.
Sendmail used to use the first entry (sort of) to determine the
host name if it could not determine from a system call easily.
Try an entry like this before the local host entry
123.123.123.123 mailserver mailserver.example.com server3
where mailserver.example.com is the public domain name for your
mail server (A record, not a CNAME)
This fixed the issue for me many yars ago on Solaris, so may
work for you now.
If all else fails, you could see if sendmail.com can help.
Good Luck
******************* REPLY SEPARATER ********************
This server has been running for years, and this is the first time a
problem of this nature has surfaced. On the other hand, the standards are
fairly clear, and the literature on Sendmail says that it should supply the
FQDN. The problem appears to be that Sendmail uses the system call
GETHOSTBYNAME, which in all Windows operating systems returns only the
unqualified host name. To get the FQDN, you must go into the registry,
extract the domain name, and append it to the host name.
I was able to get around the problem (for the time being anyway), by
manually defining the $j macro in the Sendmail.cf file to read:
Dj$w.YELLOWHEAD.COM.
How many other legitimate servers are there out there like this one? It
still seems like a dangerous assumption to make that all servers will use
the FQDN, when for so many years they have got away without having to.
J.A. Coutts