Prev: dbm map "Alias0": unsafe map file
Next: No route to host
From: Andrzej Adam Filip on 24 May 2006 08:34
perbrax(a)gmail.com writes:
> [...]
> As I mentioned, this occurs for specific messages, if I pull the
> message (qf and df file) off the queue, I can put it back later and get
> the same problem. For every problem message, we deal with a couple of
> 100000 messages without any problem, so I can't see anything major
> wrong in my configuration in general. Way down you can see the logs
> from a failed attempt to send.
>
> Anything, anyone?
> [...]

Have you tried to push such message in verbose mode?
[ using -qI... to select the message ]

sendmail -v -qI...

Full trascript of (E)SMTP session may provide more hints.

--
[pl2en: Andrew] Andrzej Adam Filip : anfi(a)priv.onet.pl : anfi(a)xl.wp.pl
http://anfi.homeunix.net/sendmail/ http://www.linkedin.com/in/andfil
Before You Ask: http://anfi.homeunix.net/sendmail/B4UAsk-Sendmail.html
From: Per Brax on 24 May 2006 08:59
Hi Andrzej,

Yeah, I did. Didn't want to make too big a post without anyone showing
any interest.
Apart from that, it doesn't say much more. Here we go, anonymized as
the previous one I hope:

Thanks for showing an interest.

Regards,

Per

# sendmail -qIBrax -v

Running /var/spool/mqinet/kBrax100000000 (sequence 1 of 1)
<Per.Brax(a)ourcompany.com>... Connecting to incoming-gw1.ourcompany.com.
via esmtprec...
220 incoming-gw1.ourcompany.com ESMTP Sendmail 8.13.6/8.13.1; Wed, 24
May 2006 10:21:37 +0200 (MEST)
>>> EHLO dmzhost1.ourcompany.com
250-incoming-gw1.ourcompany.com Hello dmzhost1.ourcompany.com
[10.83.33.90], pleased to meet you
250-ENHANCEDSTATUSCODES
250-PIPELINING
250-8BITMIME
250-SIZE 14000000
250-DSN
250-ETRN
250-DELIVERBY
250 HELP
>>> MAIL From:<per(a)nospam.brax.org> SIZE=87750
250 2.1.0 <per(a)nospam.brax.org>... Sender ok
>>> RCPT To:<Per.Brax(a)ourcompany.com>
250 2.1.5 <Per.Brax(a)ourcompany.com>... Recipient ok
>>> DATA
354 Enter mail, end with "." on a line by itself
timeout writing message to selumgw1.lu.se.tetrapak.com.: Broken pipe
<Per.Brax(a)ourcompany.com>... Connecting to incoming-gw0.ourcompany.com.
via esmtprec...
<Per.Brax(a)ourcompany.com>... Closing connection to
incoming-gw1.ourcompany.com.

From: Andrzej Adam Filip on 25 May 2006 05:13
"Per Brax" <perbrax(a)gmail.com> writes:

> Hi Andrzej,
>
> Yeah, I did. Didn't want to make too big a post without anyone showing
> any interest.
> Apart from that, it doesn't say much more. Here we go, anonymized as
> the previous one I hope:
>
> Thanks for showing an interest.
>
> Regards,
>
> Per
>
> # sendmail -qIBrax -v
>
> Running /var/spool/mqinet/kBrax100000000 (sequence 1 of 1)
> <Per.Brax(a)ourcompany.com>... Connecting to incoming-gw1.ourcompany.com.
> via esmtprec...
> 220 incoming-gw1.ourcompany.com ESMTP Sendmail 8.13.6/8.13.1; Wed, 24
> May 2006 10:21:37 +0200 (MEST)
>>>> EHLO dmzhost1.ourcompany.com
> 250-incoming-gw1.ourcompany.com Hello dmzhost1.ourcompany.com
> [10.83.33.90], pleased to meet you
> 250-ENHANCEDSTATUSCODES
> 250-PIPELINING
> 250-8BITMIME
> 250-SIZE 14000000
> 250-DSN
> 250-ETRN
> 250-DELIVERBY
> 250 HELP
>>>> MAIL From:<per(a)nospam.brax.org> SIZE=87750
> 250 2.1.0 <per(a)nospam.brax.org>... Sender ok
>>>> RCPT To:<Per.Brax(a)ourcompany.com>
> 250 2.1.5 <Per.Brax(a)ourcompany.com>... Recipient ok
>>>> DATA
> 354 Enter mail, end with "." on a line by itself
> timeout writing message to selumgw1.lu.se.tetrapak.com.: Broken pipe
> <Per.Brax(a)ourcompany.com>... Connecting to incoming-gw0.ourcompany.com.
> via esmtprec...
> <Per.Brax(a)ourcompany.com>... Closing connection to
> incoming-gw1.ourcompany.com.

I have no idea what it may be but it may help to exclude MTU related
problems described at http://www.sendmail.org/faq/section3.html#3.10

Below please find recommended test procedure (it allows to exclude
"parallel SMTP connections to target host from tracking).

0) copy sendmail.cf to sendmail-test.cf
1) in sendmail-test.cf change ClientPortOptions to contain Port
specification (comma separated list) e.g.
O ClientPortOptions=Port=1025

It will ease tracking the SMTP connection at TCP/IP packets level
2) use tcpdump (or similar tool) to track packets with local port as set
in point 1
3) Push the messages in verbose mode
sendmail -C /etc/mail/sendmail-test.cf -qI... -v

--
[pl2en: Andrew] Andrzej Adam Filip : anfi(a)priv.onet.pl : anfi(a)xl.wp.pl
http://anfi.homeunix.net/sendmail/ http://www.linkedin.com/in/andfil
Before You Ask: http://anfi.homeunix.net/sendmail/B4UAsk-Sendmail.html
From: Per Brax on 25 May 2006 16:55
Andrzej Adam Filip wrote:

> > timeout writing message to selumgw1.lu.se.tetrapak.com.: Broken pipe
> > <Per.Brax(a)ourcompany.com>... Connecting to incoming-gw0.ourcompany.com.
> > via esmtprec...
> > <Per.Brax(a)ourcompany.com>... Closing connection to
> > incoming-gw1.ourcompany.com.
>
> I have no idea what it may be but it may help to exclude MTU related
> problems described at http://www.sendmail.org/faq/section3.html#3.10
>
> Below please find recommended test procedure (it allows to exclude
> "parallel SMTP connections to target host from tracking).
>
> 0) copy sendmail.cf to sendmail-test.cf
> 1) in sendmail-test.cf change ClientPortOptions to contain Port
> specification (comma separated list) e.g.
> O ClientPortOptions=Port=1025
>
> It will ease tracking the SMTP connection at TCP/IP packets level
> 2) use tcpdump (or similar tool) to track packets with local port as set
> in point 1
> 3) Push the messages in verbose mode
> sendmail -C /etc/mail/sendmail-test.cf -qI... -v

I saw this as the last resort since there is a firewall inbetween and
also; before the upgrade I ran 8.13.1 flawlessly since it's release (or
at least since I made a Solaris package of it), so MTU problems seem a
bit far fetched. But you couldn't have known that of course, since I
didn't mention it.

But I guess I will have to do a snoop (tcpdump on Solaris) if I want to
find out what's going on, and changing the client port does of course
make my life (volumewise in information caught) much easier.

Thanks for your input.

Regards,

Per Brax

From: Andrzej Adam Filip on 25 May 2006 17:03
"Per Brax" <perbrax(a)gmail.com> writes:
> [...]
> I saw this as the last resort since there is a firewall inbetween and
> also; before the upgrade I ran 8.13.1 flawlessly since it's release (or
> at least since I made a Solaris package of it), so MTU problems seem a
> bit far fetched. But you couldn't have known that of course, since I
> didn't mention it.
>
> But I guess I will have to do a snoop (tcpdump on Solaris) if I want to
> find out what's going on, and changing the client port does of course
> make my life (volumewise in information caught) much easier.
>
> Thanks for your input.

You may check log files on the *receiving* host too before more laborious
checks. Chance that it will give a valuable hint in *this case* are low
but the check is simple.

--
[pl2en: Andrew] Andrzej Adam Filip : anfi(a)priv.onet.pl : anfi(a)xl.wp.pl
http://anfi.homeunix.net/sendmail/ http://www.linkedin.com/in/andfil
Before You Ask: http://anfi.homeunix.net/sendmail/B4UAsk-Sendmail.html
 | 
Pages: 1
Prev: dbm map "Alias0": unsafe map file
Next: No route to host