[plug] Non-existent domains with sendmail and fetchmail

[Ryan]

Howdy Plug,

I have a sendmail box which acts as a local relay at work and also
fetchmails for about 50 users every minute.  It is a temporary
arrangement while we lay plans to grab control of mail from the current
provider it is with - though it has been in place for a few years while
we do other things :)

It all works fine until fetchmail sucks down an email from a
non-existent domain:

----------
Sep 24 10:12:45 wind-tunnel sm-mta[3795]: i8O2BbrC003795:
from=<trade at expos.ae>, size=4204, class=0, nrcpts=0, proto=ESMTP,
daemon=MTA, relay=wind-tunnel [127.0.0.1]

Sep 24 10:14:12 wind-tunnel sm-mta[3856]: i8O2D3rC003856:
ruleset=check_mail, arg1=<trade at expos.ae>, relay=wind-tunnel
[127.0.0.1], reject=451 4.1.8 Domain of
sender address trade at expos.ae does not resolve
----------

It takes almost 2 minutes for sendmail to return the 4.1.8 and in this
time, the fetchmail run is in a frozen state.  It will then happen every
run as the email stays on the remote server.

With all the spam running around - this tends to happen quite a lot and
I find myself watching the fetchmail log when i get bored and manually
logging in to the remote mail host and deleting the offending messages
from which ever user accounts are affected.

Is there a fetchmail option to assist with this or a way in sendmail?  I
realise domain lookups need reasonable timeouts, but a happy medium
would be nice.  For domains it happens on a lot - I have put DISCARD
lines in the access controls for sendmail and this works, but is
obviously not a catch-all solution.

Could some RBL modifications to sendmail help at all?  I've never had
any luck getting such things to work - so if that is an answer, some
friendly pointers might be in order :)

My fetchmail lines currently look like this:

poll external.mail.host.com proto pop3 user "someusername" there with
password "password" is someusername here fetchall

Thanks,

Ryan