On Thu, 20 Mar 2003, Rob Gridley wrote: > I realize this topic has been beaten to death, but I've decided to keep on > poking it anyhow. It also might be a little OT. I wasn't able to find > anything in the list archives similar to this, so I decided to post it in > case it's useful to someone.
Not to steal your thread, but oops... Technically OT, but there's really no other place to do it. And since it seems more and more people are needing *something* to filter spam just to be on a level playing field with other providers, there's no harm in getting some SA info in the archives. Can anyone running SA on a per-user or per-domain basis speak to how this is scaling for you? Say you're running spamd on another box (or two, or three), how many users have you been able to squeeze on one box? For sake of argument, say the box is a dual-proc PIII 1GHz that is only running vpopmail/qmail handling only incoming mail. Are you getting 2,000 accounts comfortably? 5,000? 10,000? Do these numbers hold up when you're being attacked by a spammer doing a dictionary run or other evil deeds? I'm not too interested in the qmail-scanner approach, that's too much perl in the heart of the mail system. I think it would be beneficial to all involved to have some people post what type of density they can get on modern hardware. Thanks, Charles > Cheers, > > > Rob Gridley, System Administrator > MHz Design Communications Inc. > > Macs for productivity. UNIX for stability. Windows for solitaire. > > >