On Fri, Feb 20, 2004 at 06:46:45PM -0500, Michael H. Martel said: > --On Friday, February 20, 2004 5:33 PM +0100 Tomasz Kojm > <[EMAIL PROTECTED]> wrote: > > >Don't use procmail and clamav globally - such a solution is > >completely inefficient. You should scan the mail directly on the MTA > >level. > > You're suggesting something like the milter option yes ? Is it really > going to save that much/many CPU cycles and/or memory ?
Probably no more efficient than calling clamdscam out of procmail, since they both connect to a running clamd. The advantage of doing it at smtp time is that you 550 or blackhole the virus email. You don't have it stuck in your queue, you don't pass it on to potentially clueless users who ignore the giant "THIS MESSAGE CONTAINS A VIRUS!!!!" Subject:, and you don't contribute to the spread of viruses by bouncing things back after acceptance. > I've only started down this road to try and integrate it with what > we're currently using. If there's a better way (best way?) to > integrate ClamAV on a mail server, I'm open to all pointers. If you use sendmail, there's clamav-milter. If you use exim, there's the exiscan patch. I'm sure there's things available for other MTA's, but I don't use them, so I don't know about them. -- -------------------------------------------------------------------------- | Stephen Gran | This is a country where people are free | | [EMAIL PROTECTED] | to practice their religion, regardless | | http://www.lobefin.net/~steve | of race, creed, color, obesity, or | | | number of dangling keys... | --------------------------------------------------------------------------
pgp00000.pgp
Description: PGP signature
