I don't see the emails from the list.

[Bill Campbell]

On Sat, Sep 29, 2007, Jay R. Ashworth wrote:
>On Sat, Sep 29, 2007 at 06:18:00PM -0700, Bill Campbell wrote:
>> Our largest regional ISP customer only lists a single MX server
>> which handles all their incoming mail.  If there's a problem with
>> that machine, we can change the DNS to have it go to another
>> server in minutes.
>
>DNS caching client-ISP-side doesn't cause you problems there?

Not that I've ever noticed.  We're running djbdns' dnscache on the machine.
We use very aggressive DNSRBLs *AFTER* first checking an extensive DNSRBL
white list.  Since postfix doesn't support DNSRBL whitelisting, I've
enabled the tcp dictionary, and written a little tcp server in python that
handles this, caching lookups as the whitelists change very slowly.

>> various IP level checks.  It runs postfix, amavisd-new, and
>> clamav to do anti-virus and phishing checks, passing clean
>> message to a cluster of five machines that do spamassassin
>> checking and delivery to their customer's Maildir boxes.
>
>Do you have a favorite HOWTO on that?  I'm about to set up multi-domain
>email MXing on my webserver box myself.

There's no real magic here.  We have the cluster servers on a private
192.168.x.x internal LAN, using round robin DNS as poor man's load
balancing to deliver to the cluster machines.  The user's $HOME directories
are all NFS mounted and LDAP authentication.   Maildir works very nicely
with NFS as there are no locking problems.

The magic part lies in the delivery with per-user server side delivery to
IMAP folders, with spam levels and checking selectable on individual IMAP
folders.  This is all handled by code I've written.

As you may guess, all these machines are running Linux with no Windows
machines in sight.

Bill
--
INTERNET:   bill at celestial.com  Bill Campbell; Celestial Software LLC
URL: http://www.celestial.com/  PO Box 820; 6641 E. Mercer Way
FAX:            (206) 232-9186  Mercer Island, WA 98040-0820; (206) 236-1676

The difference between science and the fuzzy subjects is that science
requires reasoning while those other subjects merely require scholarship.
                -- Robert Heinlein