On Thu, 2007-11-08 at 14:47 +0000, ed wrote: > On Thu, 8 Nov 2007 15:47:53 +0100 (CET) > [EMAIL PROTECTED] wrote: > > > Thanks a lot Ed! here it sais how to setup a mail system in wich you > > can have one scanning machine and a mailbox server.. this is what I'm > > trying to do... but this won't be very helpful for me in this time > > because the mailbox server it's not qmail.. so there is no sense on > > passing him rcptto.cdb or other control files... and when I have a > > "mail hub" I have this document does but other way... my problem is > > not... how to share info between qmails to work... my main problem is > > how to handle the max incomming traffic as possible and with minimum > > machine... and I have thought that idea... > > > > thanks a lot anyway mate :) > > If you have just one box then you're screwed because you cant magic the > processor time.
Not entirely, If the main issue is timeouts during SMTP, he can move his scanning to '127.0.0.1', and remove it from his external IP. That will ensure he can receive an email from the outside in its entirety. He can throttle connections to 127.0.0.1 to prevent overload, and he won't bounce mail due to SMTP timeouts. You don't want to lose a/v scanning on your external IP, so another qmail install, with spam-only qmail-scanner, would be the cheapest solution. > If you have greater than one box the think of ways to get virus > processing away form the box that receives the mail, perhaps if you're > in an office lan, then deploy a virtual IP address protocol, so that > one of the workstations can handle the virus scanning process at that > time... might be interesting. I'd try simscan as well, it's a bit faster than qmail-scanner. But Ed really is right - the ultimate solution is more hardware. I now have a machine dedicated to only doing SpamAssassin scans... Rick !DSPAM:4733248632007426914367!
