By appending in the Posfix configuration file line smtpd_recipient_restrictions = ... spamcop and spamhaus , spam decreases in like 95% without even touching your server further (spamassassin I'm looking at you).
I you add greylisting you get rid virtually of all of spam.
I think we'd all rather "sift through a little spam than lose something I wanted" but my guess is that you've never run an email server. "a little spam" is not what you will get - it will be orders of magnitude more spam than legit emails. This is hard stuff - Gmail, supposedly one of the best, catches between 2 and 10 emails a day in my "Spam" folder that aren't actually spam. If I were to turn off the spam filter (if you could) in Gmail, I'd get 2,000 emails a day - of which 50 would be legit.
I actually have run a mail server before, and now I gladly pay someone else to do it for me :)
Point taken, but I've been on the receiving end of what I would consider false-positive blocks by Spamhaus & co. They sometimes have policies about what's considered spam that I don't think all their end users would agree with. I've been blocked for having an IP address on the same provider as someone else who allegedly advertised their website via spam. If you're running a mailing list, dealing with the anti-spam stuff is at least as big a problem as the spam it was supposed to solve.
I've never had a case of a false positive complain but yes, they can happen and I don't terribly mind because that means the sender's mail server is blacklisted and they should know.
Also in any case my recommendation is to use a 3rd party mail service, for deliverability reasons mostly.
Currently, I get a couple thousand spam messages a day, maybe 5 of wihch make it to my inbox. I haven't had a false positive in years. But if I did, I would be ok with not having to manually evaluate 50 pieces of junk for every real message I receive.
> I you add greylisting you get rid virtually of all of spam.
I don’t really like greylisting, as it takes longer for email to get through. What did help without any perceivable loss whatsoever is being extensively strict about SMTP specifications (e.g. proper hostnames in EHLO) and things like PTR records. I really like watching these ‘5xx: Client <something> rejected’ flying by in my mail.info :)
It doesn't use them to insta-spam an email, which I think is where blacklists go wrong. It gives spam points to the message if the sender is on the lists, which turns them from "blacklists" into "suspicious lists".
If an otherwise normal email happens to come from a blacklisted computer, it'll still have a chance to get through, which is the correct thing to do, in my opinion.
I still believe that everyone has the right to run their own SMTP server, and I dislike that so many places blacklist someone just because they are on a cable modem.
This sort of vague comment with no substance is just opinion stated as fact. If you have more information - hell, even a more expressed opinion, I'd be interested. As expressed, however, it's a garbage comment that waste's everyone's time. Why bother?
Not sure about "vague comment"s but anyone who has run an email server, used an rbl/rhsbl, and followed the logs <http://www.postconf.com/docs/spamrep/> would say the same. Having done so for years and run reports on dozens of servers daily it is clear that blacklists are the most effective form of spam blocking, by at least an order of magnitude.
I can't tell whether you agree with the OP or with my comment. Your first sentence argues that OP is right - "IP blacklists are a waste of everyone's time". Your second sentence though is "it is clear that blacklists are the most effective form of spam blocking, by at least an order of magnitude." Maybe a typo in your reply?
I got that part, I'm talking about your first sentence.
... anyone who has run an email server... would say the same.
This reads as though you agree that "IP blacklists are a waste of everyone's time" as OP said. And maybe you do (and that's fine) - I'm just unclear given your second sentence.
Possibly, but if Amazon isn't doing enough to secure SES against abuse by spammers(1), it's not unfair for entities intended to guard against spam to treat it as a spam source. This is one of the risks of letting third-parties run software on one's systems.
(1) keeping in mind that the definitions of "spam" can be quite subjective
Amazon SES definitely does have a spam problem, and the issue is that they don't bother doing anything about abuse reports. I've reported a spammer (who scraped our email address) to them, but the spams kept coming. Other people have reported the same thing.
If you're running a bulk mailing server you simply must respond to abuse reports, otherwise your service will get blacklisted and be essentially useless. Other providers such as mailchimp are much more proactive about getting rid of spammers.
By appending in the Posfix configuration file line smtpd_recipient_restrictions = ... spamcop and spamhaus , spam decreases in like 95% without even touching your server further (spamassassin I'm looking at you).
I you add greylisting you get rid virtually of all of spam.