[lug] A .forward standoff

Sean Reifschneider jafo at tummy.com
Tue Jun 14 16:49:54 MDT 2005


On Tue, Jun 14, 2005 at 03:30:38PM -0600, Matt Thompson wrote:
>Namely, say you have two mailservers, A and B. In A, you have a .forward
>pointing to B, and in B a .forward to A.  What occurs when you send mail
>to one?

...You get hit over the head with the LART.  :-P

If you forward an account to another account that forwards back to the
first, you have probably made a mistake.  I can't imagine a legitimate
reason you'd want to do this.  Are you trying to take down the machines
involved in doing this, as a denial-of-service exercise?  Accidents years
ago made it a required feature that MTAs detect this sort of mail loop and
as Hugh said it can be detected by a maximam hop count (basically the
number of Received header lines).  Some MTAs, notably qmail, will also add
a "Delivered-To" header as well as a Received header, and if they see a
Delivered-To line which matches the delivery they are currently trying to
do, it will break the loop without reaching maximum hop count.

Sean
-- 
 Just because something doesn't do what you planned it to do doesn't mean
 it's useless.  -- T. Edison
Sean Reifschneider, Member of Technical Staff <jafo at tummy.com>
tummy.com, ltd. - Linux Consulting since 1995: Ask me about High Availability




More information about the LUG mailing list