All,
As usual for me, understanding the problem takes 90% of the effort and
time...
At V5R3, SI17335 verifies the inbound address immediately against the SDD
rather than passing things off to MSF for later validation. This will
issue the 550 code for invalid addresses and make spamcop happy. The cover
letter is here:
http://www-1.ibm.com/support/docview.wss?uid=nas32f2a6bc7e25bf41f86256fcc004d1b37
Many thanks to all who helped. JK
On Wed May 23 16:01 , 'Rubens Lehmann' sent:
John,
I'm not on Spamcop side and not against it either, but there is a real
potencial problem:
Suppose that I forge my mail adress to use that of this list at midrange
or some other...
Then I send a message with a virus attached to a unknown account on your
server...
Then your server sends back a message to this list with that virus (as
it includes the original msg)...
Then this list server is not very well configured and resends this
message to all of us...
Then some of us have clients wich automatically loads mails attachments,
or worst, take a look at that strange file attached...
As I've said directly to you I don't know how to fix that as we don't
use OS/400 native SMTP for years, but I'm sure that there is a solution,
if not all other i5 mail based native servers would be blacklisted by
Spamcop and others, so if there's not a solution right now hope IBM
should fix it asap.
Regards,
Rubens
> -----Mensagem original-----
> De: midrange-l-bounces@xxxxxxxxxxxx
>
[midrange-l-bounces@xxxxxxxxxxxx','','','')">midrange-l-bounces@xxxxxxxxxxxx]
Em nome de johnking@xxxxxxx
> Enviada em: quarta-feira, 23 de maio de 2007 18:30
> Para: Midrange Systems Technical Discussion
> Assunto: Re: I5 smtp and spamcop
>
> All,
>
> After testing (thanks Rubens) it appears that the i5's
> SMTP server does
> not validate incoming addresses during the SMTP transaction as I'd
> originally thought. If it did, then bogus addresses would
> immediately be
> rejected with a 550 code. Instead, the server accepts the
> message, as is,
> with a 250 code and sends it to some store-n-forward
> service inside MSF.
> MSF then validates the addresses at its leisure (500
> milliseconds or so)
> and emails a "Non Delivery Report" (NDR) to the sender if
> it finds a bad
> address.
>
> This is exactly in accordance with section 6.1 of RFC2821. However
> Spamcop, being possessed of the infinite arrogance of
> those who "know
> better", has reinterpreted the RFC such that those who
> send NDRs are now
> considered spam sources and are subject to blacklisting. (See:
>
http://www.spamcop.net/fom-serve/cache/329.html).
>
> Guess it is time to open an IBM ticket. With any luck
> there is a way to
> configure the SMTP server so that it can query the SDD
> before accepting
> the message, but I'm not going to bet the farm on it. The
> alternative
> would seem to be to store NDRs internally and have our secretary
> periodically review the folder. I'm looking for volunteers
> to tell her
> that she has another thing to add to her job description...
>
> JK
>
> On Tue May 22 16:20 , Scott Klement sent:
>
> Hi John,
>
> > Chris: I'm picking my way through unfamiliar
> terminology here, but
> we're a
> > very small shop and don't have a separate mail gateway. If I
> understand
> > our config correctly, our i5 smtp server grabs
> anything that the
> router
> > sends to that particular IP address. The email is
> immediately accepted
> or
> > rejected depending on whether there is an SDD entry.
>
> Is this true? It will definitely reject mail address to
> a user with no
> SDD entry -- and it'll do it immediately (not by sending back a
> message)? Did you test that?
>
> I didn't think the SMTP server knew anything about SDD entries.
---- Msg sent via Internet America Webmail -
http://www.internetamerica.com/
midrange.com
