From owner-freebsd-hackers Sun Mar 9 23:47:47 1997 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.5/8.8.5) id XAA00962 for hackers-outgoing; Sun, 9 Mar 1997 23:47:47 -0800 (PST) Received: from who.cdrom.com (who.cdrom.com [204.216.27.3]) by freefall.freebsd.org (8.8.5/8.8.5) with ESMTP id XAA00949; Sun, 9 Mar 1997 23:47:42 -0800 (PST) Received: from misery.sdf.com (misery.sdf.com [204.244.210.193]) by who.cdrom.com (8.8.5/8.6.11) with SMTP id XAA08168 ; Sun, 9 Mar 1997 23:26:45 -0800 (PST) Received: from misery.sdf.com [204.244.213.33] by misery.sdf.com with smtp (Exim 1.59 #1) id 0w3zSM-0002ds-00; Sun, 9 Mar 1997 23:25:14 -0800 Date: Sun, 9 Mar 1997 23:25:14 -0800 (PST) From: Tom Samplonius To: Brian Tao cc: "matthew c. mead" , isp@freebsd.org, hackers@freebsd.org Subject: Re: freebsd as a news server? In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk On Sun, 9 Mar 1997, Brian Tao wrote: > On Sat, 8 Mar 1997, matthew c. mead wrote: > > > > I've got the striping factor set to 255 blocks (per a suggestion in > > the docs for ccd). > > For a news server, a stripe size of 65536 blocks (32 megabytes) > will give you optimal results, due to the mechanics of UFS's file and > directory layout in a cylinder group (which by default is 32MB). That > ccd configuration will tend to localize all disk access related to > reading a random file to a single drive in your array, thus allowing > maximum concurrency across the RAID. This is why having many smaller > disks is better than having a few larger disks. > -- > Brian Tao (BT300, taob@risc.org) > "Though this be madness, yet there is method in't" > > > But isn't choosing block and fragment sizes very important as well? The default of 8192 and 1024 seems fast, but seems to be wasting space. I remember Joe Greco mentioning something about disappointing file system performance if these are messed with the wrong way. Tom