On 18/06/15 12:22, James Triplett wrote:
On (16/06/15 18:15), [email protected] wrote:
Date: Tue, 16 Jun 2015 18:15:58 +0100
From: "[email protected]" <[email protected]>
To: [email protected]
Subject: Extended SMTP: Pipelining
Setup: OpenBSD5.7 with default OpenSMTPD
Changes: Firewall, One user, Nothing radical
Working fine:
SMTP dispatch
from second OpenBSD5.7 with OpenSMTPD (VM#2)
to receiving OpenBSD5.7 with OpenSMTPD (VM#1)
Problem:
Alien client (VxWorks-based device) won't connect
but clearly speaks to OpenSMTPD
Server Maillog message:
500-Pipelining not supported (... from OpenSMTPD...)
I had a similar error crop up when I was first installing OpenSMTPD.
So, I had to figure out what pipelining is, and why OpenSMTPD
thought it was being called for.
It turned out not be about pipelining, but a configuration issue that
eventually disappeared.
I wish I could remember the exact fix...
It may be that the 'pipelining' message is leading you in the wrong direction.
Good Luck.
James
I thought I would feed back what I found out now
that I have the VxWorks device sending e-mails TO
SMTPD (OpenSMTP).
You, James, were right - it didn't have anything to do
with pipelining in the end.
I setup a sendmail server and used -O LogLevel=15 to
capture the SMTP conversation.
During the 250-EHLO handshake response sendmail
logged that the VxWorks device had sent a QUIT -
BEFORE the end of the handshake. i.e. I couldn't
send e-mails from the device to sendmail either.
Running a manual SMTP session from another VM
revealed a 2 second or so delay in sendmail issuing the
250 EHLO response and then I realised my error.
I had deliberately prevented external access to the
DNS servers (detached the VM's WAN cable) while testing.
Reconnecting the VM WAN cable delivered an instantaneous
response from sendmail to the VM and when the VxWorks
device was reattached it connected perfectly and sent an
e-mail.
So back to OpenSMTP - ensuring VM WAN cable attached,
that /etc/hosts was properly setup and external DNS available.
BINGO!!!
It sent an e-mail to OpenSMTP too.
So it is working.
I didn't undertake a SMTP conversation log with
OpenSMTP but OpenSTMP clearly treated the
decision by the device not to await its response
during the EHLO handshake as a 500 Pipeplining error.
In fact the device had just dropped the connection
as it timed-out its send due to a long DNS lookup by
the OpenSMTP server.
Would any log information from connecting the device
to OpenSMTP in this scenario be useful to anybody????
Mike
P.S. Whatever - this may be useful to somebody else
in future.
--
You received this mail because you are subscribed to [email protected]
To unsubscribe, send a mail to: [email protected]
