[Mimedefang] (no subject)

Joseph Brennan brennan at columbia.edu
Fri Dec 5 09:17:00 EST 2008

--On Wednesday, December 3, 2008 20:56 -0200 Kenneth Irving <ken at fq.edu.uy> 

> Dear Sirs: I'm having a problem with an email that's trying to get to my
> server for the last 48 hours and I keep getting this error message
> Dec  3 20:22:33 mail sm-mta[3219]: mB3MLkaB003219: from=<xxx at x.pl>,
> size=1453865, class=0, nrcpts=1, msgid=<4936870B.1080603 at x.pl>,
> proto=ESMTP, daemon=MTA, relay=x.pl [WW.XX.YY.ZZ]
> Dec  3 20:23:35 mail sm-mta[3219]: mB3MLkaB003219: Milter (mimedefang):
> timeout before data read
> Dec  3 20:23:35 mail sm-mta[3219]: mB3MLkaB003219: Milter (mimedefang):
> to error state
> Dec  3 20:23:35 mail sm-mta[3219]: mB3MLkaB003219: Milter: data,
> reject=451 4.7.1 Please try again later
> Dec  3 20:23:35 mail sm-mta[3219]: mB3MLkaB003219: to=<zzz at fq.edu.uy>,
> delay=00:01:49, pri=30551, stat=Please try
> again later
> After little more than a minute, mimedefang returns a "timeout before data
> read". This is very odd, because the email is really passing through. I
> did a traceroute and the hole email (aprox 1.4 Mb) gets through, but then
> this timeout happens and my server is shutting down the connection
> rejecting the message.
> Someone can give me a clue about what's heppening? Thanks in advance.

I'm confused about your saying "the email is really passing through",
since sendmail is giving a 450.  The message is _not_ being delivered,

You must mean that the message is completely transmitted to your
server.  That is correct, according to the logs.  Sendmail knew the
message size, and Mimedefang is only called after the DATA phase is
complete, after the dot is sent.  I think "timeout before data read"
refers to Mimedefang processing time.

Look at the line in sendmail.cf that starts "Xmimedefang" and see what
timeouts you have in the T= equate.  That would be where the 1 minute
comes from.  Maybe if you give more time, this will just work.  But why
does it take so long?  One possibility is that you are checking RBLs
and one of them is responding very slowly or not at all.

Joseph Brennan
Columbia University Information Technology

More information about the MIMEDefang mailing list