Hello,
Recently I was asked by a client to migrate their mail services off a
legacy NT server onto a Sun workstation running Solaris 2.5.1, and I was
asked to install qmail rather than use the default sendmail. I'm new to
qmail, but I was able to use the FAQs and INSTALL docs to help me get
going fairly swiftly. (although I wish the FAQ included info about the
error message "warning: unable to open todo/xxx" -- that wasted a tonne
of time!)
About the environment. The server is a Sparc 4, running Solaris 2.5.1,
and qmail 1.0.3. The clients are all Windoze 9x/NT, using either
Netscape 4.7 or Pegasus Mail 3.1.2 as their MUA. Users use pop to send
and retrieve mail from the server The firewall is configured to
deliver incoming mail only to the mail server, and will only accept
outgoing mail from the mailserver. Host masquerading has been set up so
that all outgoing mail looks like it comes from the domain
"company.com".
There are two outstanding questions:
First, when a user on a Windows client machine uses Netscape Mail 4.7 to
send a message, the sender and return-path both say "[EMAIL PROTECTED]".
However, when using Pegasus Mail 3.1.2 to send the same message, the
return path says "[EMAIL PROTECTED]" while the sender's
address still says "[EMAIL PROTECTED]". Several remote sites refuse to
receive e-mail where the sender and return-path don't match. This
address re-writing never happened when the mail server was Exchange
running under NT, but it has become a problem since we switched to qmail
running under Solaris.
Second, the firewall people have started to complain that ever since the
switch-over to qmail, they are seeing a lot of "auth" packets to and
from the qmail server to various remote sites. They want to know what
is going on. What sort of extra packets does qmail send out? Are some
of these "auth" communications initiated by outside systems?
If I've missed something in the FAQs or docs, I apologize. Please point
me in the right direction.
Mike van der Velden
