[Mimedefang] unknown os error

[Jan Pieter Cornet]

On Thu, Dec 15, 2005 at 10:03:16AM -0600, Dave Helton wrote:
> Could someone help me decipher this error msg.
> I can ping the host by name, I can telnet to it on port 25.
> It's pretty much a valid listserver, but I just can't find the source
> of this problem.
> 
> Dec 15 09:50:39 web sendmail[26589]: jBEE79GY027808: to=<beforethebell-
> html-return-600-localuser=myserver.com at mail2.marketwatchmail.com>,
> delay=1+01:43:29, xdelay=00:00:00, mailer=esmtp, pri=2462405,
> relay=mail2.marketwatchmail.com., dsn=4.0.0, stat=Operating system error
> 
> Dec 15 09:50:39 web sendmail[26589]: jBFFjDkM026437: SYSERR(root):
> getmxrr: res_search (mail2.marketwatchmail.com.) failed with impossible
> h_errno (0)
> 
> btw, this is the only host that throws this type of rr error.
> I do not have it blocked in the firewall or in hosts.deny, or in the
> access file. I'm stumped.

This is a pretty large mail, could be the remote server is doing something
funky. The sendmail log is, as usual, less than complete and you need
insight into the real smtp conversation.

Try (on one line):
tcpdump -s 0 -w marketwatchmail.com.cap tcp and \
    host mail2.marketwatchmail.com and port 25

(optionally add "-i name_of_your_interface_to_default_gateway").

Then try sending the mail again, wait a while (until the error is
thrown and a bit more), abort the tcpdump and analyze the logfile
to see what went wrong (may I recommend ethereal and the "follow tcp
stream" option?)

My guess is something is sending a TCP packet too large for some MTU,
but the returned ICMP "fragmentation needed" is blocked... if you're not
the one blocking the ICMP packets, then you obviously cannot see that
happening.

Your only option is to send a SHORT! (less than 1000 bytes, including
headers) mail to postmaster at marketwatchmail.com and have them club their
firewall operator over the head with a large cluestick... people should
not block ICMPs at random, or Stuff Will Break.

-- 
#!perl -wpl # mmfppfmpmmpp mmpffm <pmmppfmfpppppfmmmf at fpffmm4mmmpmfpmf.ppppmf>
$p=3-2*/[^\W\dmpf_]/i;s.[a-z]{$p}.vec($f=join('',$p-1?chr(sub{$_[0]*9+$_[1]*3+
$_[2]}->(map{/p|f/i+/f/i}split//,$&)+97):qw(m p f)[map{((ord$&)%32-1)/$_%3}(9,
3,1)]),5,1)='`'lt$&;$f.eig;                                # Jan-Pieter Cornet