Re: zmailer smtp delivery problem

[Matti Aarnio <mea@utu.fi>]

> Hi,
> 
>         I've got a curious situation with zmailer and a SUN host(ecad1)
> who refuses to acknowledge requests from the deliverer(tshimshan) that
> a message is to be accepted. The smtp logfile shows a repetitive sequence
> which eventually stops when the message expires:
> ...
> 11641#  smtpopen: connecting to ecad1.phys.ualberta.ca
> 11641#  smtpconn: host = ecad1.phys.ualberta.ca
> 11641#  smtpconn: retval = 75

  This happens when connection is being attempted, none is yet made.
I wish to know what IP-address it uses, propably some silly-willy ?
As  tshimshan  is "the best MX", Zmailer throws away any MXes, and sends
with plain  A-record.  Thus it SHOULD be able to get proper address,
but one never knows.

> 11675#  smtpopen: connecting to ecad1.phys.ualberta.ca
> 11675#  smtpconn: host = ecad1.phys.ualberta.ca
> 11675#  smtpconn: retval = 75
> ...
> ecad1 is a SUN running SUNOS 4.1 and tshimshan is a NeXT running Mach2.2.
> Zmailer is running on tshimshan, and ecad1's MX record points to tshimshan.
> I can 'fake' sending a message to ecad1 on port 25 manually, and this is
> accepted and delivered.

  Me too, same thing :)  My suspect is the resolver library and pointer-
handling with  "struct hostent" -record pointers.  I had serious problems
for a while with GCC 1.x and SPARC, your problems might be of other source ?

> It looks like there is some sort of handshake failure between them, but so
> far, I've been unable to determine the problem. I've got debugging turned
> on, and for other hosts, the smtp transaction is quite verbose. But here it
> never gets started at all for ecad1.

  That is entirely within  tshimshan, I am afraid...

> I'd appreciate any comments or further suggestions as to how to proceed.
> 
> Thanks,
> -- 
> James S. MacKinnon             Office: P-139 Avahd-Bhatia Physics Lab
> Department of Physics          email : jmack@phys.ualberta.ca

	/Matti Aarnio <mea@utu.fi> <mea@nic.funet.fi>  Postmaster, ZM hacker