All these filtering schemes are like the old Christmas tree lights where if one
bulb fails, the whole thing stops working. Well sort of. I believe the RBLs can
fail, say time out, and postfix keeps working. My point though is you need to
consider the possibility of the mail server going down due to too many tools in
the chain.
I run on a VPS. I have a single point of failure. I manage the server and
really don't want to drop everything to fix a clogged email queue. (Amavisd
would do that to me.) I may not have a computer handy.
You will never achieve spam blocking perfection, and false positives are an
issue. I would just mark the email as spam when I ran spamassassin, so I ended
up looking at the spam email anyway.
Original Message
From: ph...@caerllewys.net
Sent: April 23, 2019 11:50 AM
To: postfix-users@postfix.org
Subject: Re: spam from own email address
On 4/23/19 2:40 PM, lists wrote:
> I would investigate using rspamd rather than spamassassin. At the moment
> I run neither since I have settled upon a nice mix of RBLs and check the
> reverse pointer. That Perl code to get rid of dynamic domains really
> helps nuke spammers.
>
> Spamassassin tends to use a lot of memory. When I was using it, I had it
> on a rather memory limited VPS and actually needed to use VM.
I'm using rspamd myself, though it's clear I still have a lot to learn
about configuring and training it. I used to use DSpam, and was getting
excellent results with it — something over 99.97% accuracy — but it was
abandoned and eventually became unmaintainable.
--
Phil Stracchino
Babylon Communications
ph...@caerllewys.net
p...@co.ordinate.org
Landline: +1.603.293.8485
Mobile: +1.603.998.6958
