Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 5 Jan 2017 09:06:48 -0600 (CST)
From:      "Valeri Galtsev" <galtsev@kicp.uchicago.edu>
To:        "Matthias Fechner" <idefix@fechner.net>
Cc:        freebsd-questions@freebsd.org
Subject:   Re: Filtering Email
Message-ID:  <46452.128.135.52.6.1483628808.squirrel@cosmo.uchicago.edu>
In-Reply-To: <99370884-b547-c3c7-7b2b-711cedcedf27@fechner.net>
References:  <2E557AFF-35A1-4D08-8FA9-10C65BF4ABDE@lafn.org>    <9bd488a3-ca45-a546-3706-3b032386f954@FreeBSD.org>    <99370884-b547-c3c7-7b2b-711cedcedf27@fechner.net>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, January 5, 2017 2:58 am, Matthias Fechner wrote:
> Am 05.01.2017 um 09:12 schrieb Matthew Seaman:
>> There's a potential problem with rejecting email from your local server
-- backscatter.  If your upstream MTA has accepted a message for
delivery and then your local MTA later decides to bounce it, there is
no
>> choice other than to send the bounce to the sender address in the mail
headers, and spammers nowadays forge that address, so you end up
resending the spam to some (possibly innocent) third party.  It's
better
>> to just /dev/null the messages in such circumstances.
>
> hm, this is really dangerous, you could kill by this action real email
without notify the sender.
> I would recommend that the sender set a SPF record to control which
server is allowed to use their domain.
> This would mitigate this problem and forge of email headers is not
possible anymore.

I agree both with both Matthew and Matthias. It is bad to drop messages
your server accepted - you may have to answer one day where is message
with messageID "this". And it is really nasty to make upstream server a
source of backscatter, so their admins will be same nasty to you in
return. One of examples (many may be surprised) is gmail: they accept the
message addressed to [somebody, whatever is here]@gmail.com no matter
whether accounts exists or not, and if not, they come back to you with
nondelivery notification. This sets me off as on my server if I have to
forward message, then before accepting it I check downstream if it is
deliverable on destination server, if not I'm just not accepting the
message. But gmail how they operate potentially can make my server a
source of backscatter (yes, I had an incident!), so I don't forward to
gmail.

Back to the subject. The best approach to deal with spam messages we found
for ourselves is: label them as spam, and deliver them into separate
mailbox of recipient. We avoid using Junk mailbox, as this is often used
by client software (Thunderbird, Apple Mail, etc.), using mailbox with
different name simplifies distinguishing what detected message as spam:
your server or client software.

Analyzing messages: look at spamassassin, or amavisd (the last uses
spamassassin, and can harness much more). Mail is rather big department of
system administration, and dealing with spam has significant share in it.
Be ready to spend some time, and re-visit it every so often.

I hope, this helps.

Valeri

PS I'm not certain gmail still behaves the way it did in the past - as I
described it. It did this weird thing in the past. That time someone
joked: they try to collect all information that comes their way (for
whatever profit), that's an explanation. I'm reluctant to re-visit my
attitude to bad things of this world that I had to deal with ;-)

>
>
> Gruß
> Matthias
>
> --
>
> "Programming today is a race between software engineers striving to
build bigger and better idiot-proof programs, and the universe trying to
produce bigger and better idiots. So far, the universe is winning." --
Rich Cook
>
>
>


++++++++++++++++++++++++++++++++++++++++
Valeri Galtsev
Sr System Administrator
Department of Astronomy and Astrophysics
Kavli Institute for Cosmological Physics
University of Chicago
Phone: 773-702-4247
++++++++++++++++++++++++++++++++++++++++







Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?46452.128.135.52.6.1483628808.squirrel>