SUMMARY: sendmail errors: collect: premature EOM: Connect reset by (remote host)

Thanks to the following people for replying:
Shannon Ward <sward@rackspace.com>
mcferren@colltech.com
Landon Stewart <landons@uniserve.com>


Shannon Ward <sward@rackspace.com> seems to have hit the nail
right on the head!  Thank you, Shannon!


If you are using a Cisco firewall product that supports their fixup
protocol,
try disabling it for smtp.  As far as I know, it only
supports SMTP and you may be using ESMTP.

--Shannon Ward


Turns out that after our telecomm guy adjusted the Cisco Pix,
the errors were dramatically reduced.  Now instead of 2 or 3
of those errors per minute, I get one error per hour or two.


------------------------------------------------
Original question:


I'm running Solaris 2.6 and Sendmail 8.10.2.  I have two mail
servers which accept incoming mail.

On both hosts, the following error messages repeats quite frequently:

Aug 27 17:04:15 nserver1 sendmail[913]: f7RL3Ax00913: collect: premature
EOM: Connection reset by exchange.shagmail.com
Aug 27 17:04:15 nserver1 sendmail[913]: f7RL3Ax00913: SYSERR(root):
collect: I/O error on connection from exchange.shagmail.com,
from=<CustomerService@getyourfreebies.com>: Connection reset by
exchange.shagmail.com


I've looked at Sendmail's FAQ 3.10, but I doubt that it's the MTU
setting (this is a dedicated non-SLIP/PPP connection to the Net) or
ICMP packets being blocked since the mail servers are pingable from
the Net.

Anyone have any ideas what this could mean?

There seems to be a lot of people asking this on the Net, but very
few intelligent replies (and certainly no solutions to this problem).


Any suggestions would be greatly appreciated.

I will summarize.

Thanks in advance,
Jeff