On Wed, 2005-11-30 at 08:33 -0800, Robert wrote:
} } Ugh.. qmail. I'll refrain from spitting up too much hate about the DJB } way of doing things here and just say "it's a functional product, but } his way is not the right way" } } There are some excellent mail setup guides located at } http://www.hughesjr.com/content/category/1/14/30/Guides } } They're quite good, and cover everything from intial setup to webmail } to filtering. Well worth the read. } Jim Perrin
thanks for input!!!
having built many mailservers i will be ok in general yet what i am really looking for now is how is everyone here filtering "spam in languages other than English"... most of the spam i get is in Spanish.
---- - Google offers to translate pages with a sublink on all found items that aren't your primary languages...at least it does for me.
- spam is spam and is not about language. If you want more aggressive spam filtering, perhaps you should investigate a better wrapper (I use MailScanner but my guess is that mime-defang and amvaisd-new are quite capable), implementing more spam features (DCC, pyzor, DNSBL, etc.)
- greylisting (http://www.greylisting.org)
I use postfix, greylisting, MailScanner with spamassassin & clamav, and sieve filters (cyrus-imapd) and get pretty good control. Probably the most significant things I do is to have a fairly comprehensive smtp_helo/smtp_sender checks in postfix and greylisting which gets all the low hanging fruit so I am not sucking massive cpu power on the rest of the checks mandated by MailScanner ----
i am having a hard time finding it googling (other than Bayesian filters) and i just need some searching help or direction to find so i can study up please?
---- http://wiki.apache.org/spamassassin/ ----
and thing is, ummmm im willing to use sendmail yet i want my customers to be able to admin their own virtual mailservers with a web interface... what is everyone using there??? sendmail and ??? or qmail and ???
---- on servers that provide smtp for domains...probably 60% sendmail, 35% postfix, 5% other - just a wild guess - no empirical data whatsoever.
Craig