18 messages in net.sourceforge.lists.courier-maildropRe: [maildropl] Re: Even more flock()...
FromSent OnAttachments
Mitchell VincentApr 29, 2004 10:19 am 
Ron JohnsonApr 29, 2004 11:21 am 
Mitchell VincentApr 29, 2004 12:21 pm 
Ron JohnsonApr 29, 2004 12:53 pm 
Mitchell VincentApr 29, 2004 1:19 pm 
Sam VarshavchikApr 29, 2004 5:36 pm 
Mitchell VincentApr 30, 2004 4:38 am 
Ron JohnsonApr 30, 2004 5:48 am 
Mitchell VincentApr 30, 2004 6:48 am 
Tony EarnshawApr 30, 2004 7:21 am 
Mitchell VincentApr 30, 2004 7:35 am 
Devin RubiaApr 30, 2004 9:48 am 
Tony EarnshawApr 30, 2004 12:06 pm 
Mitchell VincentApr 30, 2004 12:16 pm 
Ron JohnsonApr 30, 2004 3:34 pm 
Sam VarshavchikApr 30, 2004 4:01 pm 
Tony EarnshawApr 30, 2004 5:36 pm 
Ron JohnsonApr 30, 2004 7:33 pm 
Actions with this message:
Paste this link in email or IM:
Paste this link in email or IM:
Atom feed for this thread
Paste this URL into your reader:
Subject:Re: [maildropl] Re: Even more flock() problems..Actions...
From:Ron Johnson (ron.@cox.net)
Date:Apr 30, 2004 3:34:45 pm
List:net.sourceforge.lists.courier-maildrop

On Fri, 2004-04-30 at 15:16 -0400, Mitchell Vincent wrote:

Some pretty large organizations are doing this, for example (I'm a Postfix person, was Exim 4 and it's more or less the same there) with amavisd-new and SA. They even run anti-virus shit on them. They do it by throwing high-RAM and high I/O disk hardware at it. I believe Brian Candler might be one. [...]

If I could get "him" to spend a little bit of money we wouldn't have to worry..

The disk IO is where the bottleneck is.. RAID 5 on UW SCSI hard drives - it's just not cutting it anymore between all the IMAP/POP3 and SMTP traffic...

What about an intelligent, caching controller? 512MB of battery- backed-up RAM on the controller will make RAID 5 sing. We have dual-redundant controllers, each with 512B RAM, and that *really* smokes.

But then, when do you get the time to migrate over to it......

-- Ron Johnson <ron.@cox.net>