From owner-freebsd-hackers Sat Apr 29 10:44:20 1995 Return-Path: hackers-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id KAA00867 for hackers-outgoing; Sat, 29 Apr 1995 10:44:20 -0700 Received: from ns1.win.net (ns1.win.net [204.215.209.3]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id KAA00860 for ; Sat, 29 Apr 1995 10:44:17 -0700 Received: (from bugs@localhost) by ns1.win.net (8.6.11/8.6.9) id NAA27122 for hackers@freebsd.org; Sat, 29 Apr 1995 13:46:01 -0400 From: Mark Hittinger Message-Id: <199504291746.NAA27122@ns1.win.net> Subject: multi-pci scsi controllers To: hackers@FreeBSD.org Date: Sat, 29 Apr 1995 13:46:00 -0400 (EDT) X-Mailer: ELM [version 2.4 PL23] Content-Type: text Content-Length: 1245 Sender: hackers-owner@FreeBSD.org Precedence: bulk I read Jordan's note about building a large server with PC hardware with keen attention. We are doing the same here. I'm interested in the issue of using multiple PCI scsi controllers for disks. On my main server I've got enough disk that I'm starting to think that I do have some controller contention. How much does a 2940 or bt946c burden a PCI bus? I'm sure that the walnut creek guys already know these details. I have a handfull of these controllers and I could experiment with spreading some load around. Any caveats with multiple 2940's that we should be aware of? Could we mix and match 2940/bt946c in the same box without bios contention? PS: we got the Kentucky Derby home page business on our main server so the Derby home page is on FreeBSD! http://www.win.net/derby Lots of hits - and the first saturday in may is coming. I worked at another horse racing database firm in my younger days - and derby eve the machines were swamped with activity from all over the world. We are getting enough news and web activity on the server to probably justify two controllers at this point - if it would make a positive difference. I probably need to scare up some more RAM (as usual). Regards, Mark Hittinger bugs@win.net