SpamD and greylisting in general?
Most of the spam filters I've encountered in the past "learn" from what the user says is spam. I'm curious as to how spamD and greylisting in general works. Does the sender get an annoying "could not deliver, will try again later" email when the recipient uses greylisting (spamD in particular) or does the sending server just hang onto the mail and try again a few minutes later? Basically, I'd like the spam filter to be completely transparent to the sender. Delivery times I understand will improve after the filter learns what is good and what isn't so that doesn't bother me. Is spamD right for me?
It's transparent to the user. Spamd will basically just tell the incoming mail server to go away for a little while… and theoretically only the real mail servers will come back. A typical spammer will not try again.
I've never used spamd, but most greylisting packages will remember which servers have previously sent messages, so the delay is only in the first email received.
I have my server only greylist incoming mail senders that have dynamic ips. Very few real mail servers have dynamic ips, and very few spammers send spam from static ips. This has the added benefit of far fewer lost emails (which are rare anyways), and no delay for 99% of incoming mail.