what could cause failure to print sent.c's "to=<...>" line?

[Meng Weng Wong]

I am trying to troubleshoot a very rare syndrome: a
successful smtp delivery is normal in every respect except
that the "to=<...>" line does not appear in the logs. This
causes muttering and much consternation.

Oct 1 18:32:43 smtpd[5423]: D0E793E640: client=localhost.localdomain[127.0.0.1]
Oct 1 18:32:44 cleanup[4905]: D0E793E640: message-id=<004501c14ac7$78b4d9b0$1200...@gsicomp.on.ca>
Oct 1 18:32:44 qmgr[4884]: D0E793E640: moving from incoming -> active
Oct 1 18:32:44 qmgr[4884]: D0E793E640: from=<owner-freeb...@FreeBSD.ORG>, size=3551, nrcpt=1 (queue active)
Oct 1 18:32:44 qmgr[4884]: D0E793E640: deleting queue file under defer
Oct 1 18:32:44 qmgr[4884]: D0E793E640: assigning recipients to delivery requests and closing messagefile.
Oct 1 18:32:44 qmgr[4884]: D0E793E640: qmgr_deliver: attempting send request to transport=smtp, recipient=ham...@execpc.com
Oct 1 18:32:44 qmgr[4884]: D0E793E640: qmgr_deliver finished writing to transport=smtp, recipient=ham...@execpc.com
Oct 1 18:32:44 smtp[5398]: D0E793E640: handling request from owner-freeb...@FreeBSD.ORG to ham...@execpc.com
Oct 1 18:32:45 smtp[5398]: D0E793E640: communicating status 0 back to qmgr.
Oct 1 18:32:45 qmgr[4884]: D0E793E640: qmgr_active_done starting
Oct 1 18:32:45 qmgr[4884]: D0E793E640: deleting queue file under active
Oct 1 18:32:45 smtp[5398]: D0E793E640: communicated status back to qmgr with errorcode 0. responsibility ends.
Oct 1 18:32:45 qmgr[4884]: D0E793E640: all done. removed.

I have been observing something like this roughly once in
every 100,000 deliveries.

Could tainted external input (eg. in the 250 response status
text) cause the msg libraries to decline to print a msg_info()?

I'm trying to rule out as much as possible before I blame
syslogd. Under what conditions could the msg, vstring, and
syslog libraries conspire to not print a line of output?
Maybe the smtp server response is exceeding some buffer
length? Any thoughts?

-
To unsubscribe, send mail to majo...@postfix.org with content
(not subject): unsubscribe postfix-users

[Ralf Hildebrandt]

On 2 Oct 2001 06:01:40 +0200, Meng Weng Wong <meng...@dumbo.pobox.com> wrote:
> I am trying to troubleshoot a very rare syndrome: a
> successful smtp delivery is normal in every respect except
> that the "to=<...>" line does not appear in the logs. This
> causes muttering and much consternation.

Maybe it's lost by syslogd due to load/bandwidth limitations?

--
ralf.hil...@innominate.com innominate AG
Technical Consultant Don't be afraid of what you see -
Diplom-Informatiker be afraid of what you don't see!
tel: +49.(0)7000.POSTFIX fax: +49.(0)30.308806-77

[Meng Weng Wong]

On Tue, Oct 02, 2001 at 06:00:00AM +0000, Ralf Hildebrandt wrote:
|
| Maybe it's lost by syslogd due to load/bandwidth limitations?
|

yes, since my last post i managed to gather enough evidence
to blame syslog --- logging goes over udp so some packets
are being lost. i will probably upgrade to syslog-ng and
use its tcp facility. no worries, postfix is not losing mail.

meng

[Ralf Hildebrandt]

On Tue, Oct 02, 2001 at 10:41:11AM -0400, Meng Weng Wong wrote:

> yes, since my last post i managed to gather enough evidence
> to blame syslog --- logging goes over udp so some packets
> are being lost. i will probably upgrade to syslog-ng and
> use its tcp facility. no worries, postfix is not losing mail.

One line in 100.000. OK, syslog seems to suck under high loads, but
if you use TCP, you could saturate your network with logging alone
(unless you use a dedicated connection to a loghost)

--
Ralf Hildebrandt http://www.arschkrebs.de
Treat your password like your toothbrush. Don't let anybody else use it,
and get a new one every six months. --Clifford Stoll