Re: [Exim] rejected recipients

Top Pagina
Delete this message
Reply to this message
Auteur: Dave C.
Datum:  
Aan: Philip Hazel
CC: Salvatore Greco, exim-users
Onderwerp: Re: [Exim] rejected recipients
On Tue, 22 Aug 2000, Philip Hazel wrote:

> On Tue, 22 Aug 2000, Salvatore Greco wrote:
>
> > Any ideas why my exim 3.16 MTA is rejecting email from this address:
> >
> > 2000-08-22 01:36:03 reject all recipients: 3 times bad sender <f1.sasha.co.za@???> H=www.i
> > nfonetix.co.za (infosrv1.infonetix.co.za) [196.25.201.76]
>
> You have sender_verify set. The client sent
>
> MAIL FROM:<f1.sasha.co.za@???>
>
> infosrv1.infonetix.co.za is not registered in the DNS. Therefore,
> f1.sasha.co.za@??? is a "bad sender". The sending
> host has sent it to you 3 times within 24 hours:
>
> (1) The first time, Exim rejected the message after the DATA phase; it
> will have logged the message's headers in rejectlog.
>
> (2) When the client didn't pay attention to the permanent error code,
> (in contravention of RFC 821) and sent the message again, Exim rejected
> the MAIL FROM command with a permanent error.


It is of course possible that the same sender-host is sending a
different message, with the same sender? That doesn't excuse it of
course, nor make the message(s) any more acceptable, but it is
possible, is it not?

> (3) The very broken client tried a third time to send the message, and
> this time Exim rejected every RCPT TO command, in the hope that the
> client at least recognizes those rejections as meaning "I'm never going
> to accept this message. Go away. Return it to sender."


Consider the case of someone trying to send out a big pile of batches
of spam, either trying to use you as a relay, or sending messages to
addresses in your local domain(s) directly to your server. While it
would certainly still be desirable to block invalid senders and/or
spam, which often go hand-in-hand anyway, it might not be a case of a
broken SMTP client, it might be just trying to deliver each batch,
which all have the same (invalid) sender-address.