From owner-freebsd-questions Fri Apr 11 16:28:09 1997 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.5/8.8.5) id QAA29476 for questions-outgoing; Fri, 11 Apr 1997 16:28:09 -0700 (PDT) Received: from mixcom.mixcom.com (mixcom.mixcom.com [198.137.186.100]) by freefall.freebsd.org (8.8.5/8.8.5) with SMTP id QAA29465 for ; Fri, 11 Apr 1997 16:28:05 -0700 (PDT) Received: by mixcom.mixcom.com (8.6.12/2.2) id SAA00978; Fri, 11 Apr 1997 18:28:14 -0500 Received: from p75.mixcom.com(198.137.186.25) by mixcom.mixcom.com via smap (V1.3) id sma000922; Fri Apr 11 23:27:51 1997 Message-Id: <3.0.32.19970411182037.00b64620@mixcom.com> X-Sender: sysop@mixcom.com X-Mailer: Windows Eudora Pro Version 3.0 (32) Date: Fri, 11 Apr 1997 18:20:38 -0500 To: Anthony Barlow From: "Jeffrey J. Mountin" Subject: Re: AMD 586 CPU Cc: freebsd-questions@freebsd.org Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Sender: owner-questions@freebsd.org X-Loop: FreeBSD.org Precedence: bulk At 10:03 AM 4/10/97 +0100, Anthony Barlow wrote: >We're building a AMD 586 150. The machine is planed to be a mail server >accepting both inbound and outbound mail, it will also be our POP server as >well as run 4 light mailing lists. As it's a 'major server' you can >appreciate why we've asked. AT the moment our mail server is an Intel P90 >running Linux 1.2.13 and as performed flawlessly with uptimes in excess of >60 days, it's never crashed, we only reboot it to clear the RAM and any >possible stuck processes. The reason for the change is that during peak >periods it will see loads of around 8, however 99% of the time the load is >0.01, where wondering with the price of AMD 'pentium' chips, if we could >the load would reduce during those heavy periods. Why a 150? We are talking about a server with a lot of disk IO and the PCI bus for a 150 is 30, which is why I stay with 100/133/166/200 CPU speeds. Also good to use several smaller drives. >BTW we know what causes the load - procmail when it delivers our >announcement mailing list. I've just seen in a previous message something >about qpoppers bullitin feature, I wonder if using that feature will get >around this problem? Never used procmail, but it sounds like it should have a CPU limit flag, just like sendmail does, so when load high it doesn't spawn. I heard in qpop it was broken, so I imagine that I'll be doing a script that will use a faked header (so POP can read it), lock the mailbox, copy the message, etc. With this my ability to choose which users/groups will receive a copy is greatly expanded and is faster than using mail with less disk IO. ------------------------------------------- Jeff Mountin - System/Network Administrator jeff@mixcom.net MIX Communications Serving the Internet since 1990