On Friday 13 January 2012 16:57:21 Al Zick wrote: > On Jan 12, 2012, at 3:57 AM, Egoitz Aurrekoetxea Aurre wrote: > > Apart from this if you use some trustable RBL, perhaps ------------------------------------^^^^^^^^^^^^^ > > greylisting and you update Spamassassin rules regularly... > > you should be pretty fine.. > > I am not using spamassassin, or greylisting, and I just removed > the RBL's because of waaaaaay too many false positives.
This is absurd. It surely means you chose some overly-aggressive (not trustable) DNSBLs. To say that all DNSBLs cause loss of mail is ridiculous. Why didn't you even consider it worth mentioning what lists you were using? Yes, if you followed some silly outdated howto written by someone who didn't even understand it way back then, you get what you deserve. For general use, both Spamhaus Zen and Barracuda BRBL are safe and effective. Spam-eating monkey is moving up into that list for me. Always ALWAYS know the policies of any DNSBL you are considering. http://www.postfix.org/postconf.5.html#warn_if_reject is your friend; you can take a DNSBL out for a trial run without risking anything. > Right now most of my filtering is being done with a set of > rules for procmail and with bogofilter, although I am > considering using policyd and dovecot with sieve plugin. > > Policyd (or graylisting) should mean more email is rejected I do not recommend greylisting now, especially not if using a good DNSBL to catch the bots who retry. Greylisting will still take out quite a lot of the bot spam, but it's far less effective than it was before the ratware pushers adapted to it. All I can say is that I'm doing pretty well with my postscreen. Sometimes we see a few spams & 419s from hotmail or gmail, but the direct-to-MX botnets are not getting through very often. > (like it should be) and a filtering during dovecot should mean > that whitelisted servers still get their email filtered. -- http://rob0.nodns4.us/ -- system administration and consulting Offlist GMX mail is seen only if "/dev/rob0" is in the Subject:
