Afterthought about bouncing

Stephe Campbell campbell at CNPAPERS.COM
Fri Feb 13 14:50:51 GMT 2004

Mr. Kercher & Mr. Earickson,

Sorry to be so loose-worded in my query. My main problem is I have always
taken the easy way of configuring sendmail with linuxconf. I inherited a
mail server configured this way and never changed it. Now, with all of this
spam crap, I need to use, or at least think I need to use, more of what
sendmail offers than what linuxconf would do for me.

Based on what both of you have said, I may not have to change over. I have
looked many times for a way to drop messages completely at the MTA based on
whether the user exists or not, similar to the way a firewall can drop
packets. But it looks like at the very least, I will always either generate
and send a "user unknown message" back to the sender (real or not) or
forward the message to this catch all (dead box) account. Doublebounce would
handle the return from the return if one is generated.

If I don't use the catch all account, though, am I really returning the
email back to the sender, or just the notification that the user doesn't
exist and dropping the original email? Either way, aren't I spamming the
innocent returnee who probably never sent the original? Unless I misread Mr.
Kercher below, he indicates that my server will not generate an email to the
*sender*, but won't it still send the "user_unknown" reply back. How do I
stop the "user_unknown" reply other than using a catch all account?

Thanks very much. Does this make sense?

Steve Campbell
campbell at
Charleston Newspapers

----- Original Message -----
From: "Mike Kercher" <mike at CAMAROSS.NET>
Sent: Friday, February 13, 2004 8:56 AM
Subject: Re: Afterthought about bouncing

> I'm not sure I understand exactly what you're wanting to do, but I'll give
> it a shot.
> If an email is not deliverable due to an addressing error, your machine
> should reject the email connection with a "user unknown".  Your server
> not generate an email to the *sender*.  Their MTA would bounce the message
> back to the sender as undeliverable.  This is not any additional load on
> your server.  I would leave it up to the sender to make sure the address
> correct before sending...not the postmaster.
> If you want to catch all of the mis-addressed emails, add an entry to your
> /etc/mail/virtusertable:
> localuser at
> This is called a catch-all.  You could also add a few entries for commonly
> mispelled addresses such as:
> johndeo at      johndoe at
> m.smith at      msmith at
> Mike
> > -----Original Message-----
> > From: MailScanner mailing list
> > [mailto:MAILSCANNER at JISCMAIL.AC.UK] On Behalf Of Steve Campbell
> > Sent: Thursday, February 12, 2004 8:49 PM
> > Subject: Afterthought about bouncing
> >
> > To the list,
> >
> > I really never thought about this until my recent explosion
> > of invalid-user emails sent to our domains and my
> > undetermined reason for slowness. Could someone explain the
> > definition of bouncing to me in their own environment and
> > tell me what they think of the below stuff?
> >
> > I use a dead box here to capture nonexistent user emails,
> > where they are reviewed for simple mis-typed addressing, and
> > then either forwarded to the proper user or discarded. I
> > blacklist heavily on "From:" to throw away emails.
> > In this situation, I hope I am not bouncing much.
> >
> > As I research my move to using M4 for pure Sendmail, getting
> > away from linuxconf, I have not seen a way to prevent the NDR
> > message generated by sendmail when I do not use a fallover
> > address (dead box). It appears that this will always occur.
> > Isn't this a form of bouncing even though I am not forwarding
> > the original message? I want to eliminate as much for MS to
> > do as possible, and want to do this at the MTA (SendMail, again).
> >
> > Please no wars about bouncing. Just the facts, ma'am. (That's
> > from the TV show Dragnet to all of you youngsters).
> >
> > Thanks,
> >
> > Steve Campbell
> > Charleston Newspapers
> > campbell at
> >
> >
> >
> >
> >
> > -------------------------------------------------
> > This mail sent through IMP:
> >

More information about the MailScanner mailing list