[Mimedefang] Truncated connections causing multiple resends on incoming mail

[Kimmo Jaskari]

I'm having an issue, related to mails that are large-ish (5-10MB
attachment or so, perhaps) and I'm not really sure where I should tackle
this issue.

This is on Sendmail 8.13.7 on Solaris 10 (standard Sun-compiled binary,
custom cf file) and MIMEDefang 2.59. Using SA and ClamAV (daemon) to
scan.

I'm assuming it is some kind of timeout issue, but I've set both the
MIMEDefang timeout (in the init script) to 29 minutes and the milter
timeout in the sendmail mc (and cf) to 30 minutes, and even though that
strikes me as very high that seemed to have no effect on this.

What happens is that the process of receiving mail cuts out before the
mail is completely transferred, which causes the server talking with
mine to resend... and resend... and resend... and the recipient on my
end gets mail after mail. Essentially, the remote server sees this part
of the process: 

[opens connection]	
	220 server.example.net Greetings
HELO client.example.com	
	250 Welcome client.example.com
MAIL FROM: <joe at client.example.com>	
	250 joe at client.example.com... Sender ok
RCPT TO: <jane at server.example.net>	
	250 jane at server.example.net... Recipient ok
DATA	
	354 Enter mail

... and that's it, the connection seems to get cut off there. At this
point, when the mail has been transferred, they should see:

[transmits message]	
	250 Message accepted
QUIT	
	[closes connection]

But that doesn't seem to happen, so the remote server gets confused as
to whether or not it got through and keeps it in the queue and resends.

Does anyone have any good ideas as to where I could start looking at
what could be causing this? Thanks!

/Kimmo