Re: debian-user list info and guidelines: spam
On Mon, 25 Jan 2021 08:55:53 +0200
Andrei POPESCU <andreimpopescu@gmail.com> wrote:
> On Lu, 25 ian 21, 01:03:07, John Kaufmann wrote:
> > > In the last seven days we've seen bounces for the following list:
> > > * debian-user
> > > 1 bounce out of <nn> mails in one day (<x>%, kick-score
> > > is 80%)
> >
> > First: How common is this occurrence for others?
>
> Depends on the provider. I receive list mail to a GMX account and it
> does happen from time to time.
>
> > Even when it happens, it will obviously never rise to the
> > kick-score,
>
> It could happen quickly with a misconfigured e-mail server / spam
> filter, which is why the warning is being sent out.
>
Or with a mail server that's down for a while, or where the Internet
connection is down, and in my case at least, when Debian's DNS servers
do not respond in a reasonable time.
> > so it's functionally forgettable. But the post itself is
> > educational, particularly with the background info links, which
> > explain the origin of the problem in spam control, and why bouncing
> > list mail is not nice, and what can and should be done about spam.
> > I'm sympathetic to the problem, but don't run my own mail server [I
> > suspect many on this list do so, and thus are closer to spam
> > issues], and can never get my ISP to care about this.
I run my own mail server, which does *not* reject SMTP transactions
based on content, it looks for respectability. It never rejects Debian
email, whether spam or not.
>
> We can at least pester them about it. If enough of us do, eventually
> they will have no choice but to listen, especially in case of paying
> customers.
I have a client who uses BT as his ISP, and therefore must use
Microsoft as his email provider. It is not possible to affect their
spam policy in any way, or even to turn it off.
> >
> > Is there an intelligent way to manage when spam control efforts
> > break the system they want to protect? Do the debian lists attempt
> > to control spam by blocking specific users, domains or ISPs?
I've never found content-based anti-spam methods to be reliable. For a
time, I tried spamassassin, but I found I was spending a lot of time
tweaking it, and a week later would need to do it again. I decided that
spammers evolve faster than I care to try to deal with.
I do segregate some emails that cannot possibly be anything other than
spam but do not reject them, and send myself a daily email listing the
very few that were rejected that were actually sent to a real user on
my domains. I do get the occasional incorrect rejection, but no more
than half a dozen a year, which I can normally put right the next day.
--
Joe
Reply to: