Oliver Fromme wrote: > > Personally, I tend to prefer false negatives over false > positives. After all, greylisting is not the only measure > against spam, and if something gets through, the next > stage filter (e.g. content-based or whatever) can take > care of it. YMMV, of course. > > That was my thinking when I set mine spam stuff up years ago and found out about greylisting. My set up is: Pregreet, sendmail.mc setting to get rid of ratware programs running on botnets. greylist, selfexplaining here. ;-) clamav, milter-calmav to take care of the viris and phising mails. spamassassasn, milter-spamassassan takes care of the ones that slip through and is the last due to most resource one of them all and is set up to flag/tag SPAM in the subject line only. And reject for very high scoreing ones. This setup has no false positves except for spammassassan and it can only tag it with SPAM and send it on its way. If there is a technicial glitch or whatever and letgit mail gets rejected, the sender ACTUALLY gets an error message and not to /dev/null. So the sender can use another channel, such as IM, IRC, phone, etc. to alert the sysdem admin of a problem. Its a lot easer to whitlist problem servers then to blacklist all the spammer servers out there. --Techwolf.
Message
Re: [milter-greylist] BotNet plugin
2007-01-05 by Techwolf
Attachments
- No local attachments were found for this message.