From owner-freebsd-hackers Wed Mar 4 03:40:49 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id DAA16650 for freebsd-hackers-outgoing; Wed, 4 Mar 1998 03:40:49 -0800 (PST) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from bsd.synx.com (rt.synx.com [194.167.81.239]) by hub.freebsd.org (8.8.8/8.8.8) with SMTP id DAA16621 for ; Wed, 4 Mar 1998 03:40:23 -0800 (PST) (envelope-from remy@synx.com) Received: from s3.synx.com (s3 [192.1.1.247]) by bsd.synx.com (8.6.12/8.6.12) with SMTP id MAA21616; Wed, 4 Mar 1998 12:44:52 +0100 Received: from rs1 by s3.synx.com id aa25652; 4 Mar 98 12:30 GMT Date: Wed, 4 Mar 1998 11:29:28 -0100 (GMT) From: Remy NONNENMACHER To: Terry Lambert cc: hackers@FreeBSD.ORG Subject: Re: Disk write caches In-Reply-To: <199803040313.UAA12989@usr05.primenet.com> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Wed, 4 Mar 1998, Terry Lambert wrote: > I would be very interested to see what the benchmarks on a machine > with soft updates would be with a disk with and without write > cacheing. > I tested cache benefits within disks. Write cache disabling lowered the perfs from 2 to 10 time depending on operations. RAIDying inside the kernel can be a good 'cheap' solution even if i prefer it external (for the same reasons as Mr Shapiro). What i humbly demand is that, for those able to plug an UPS on their machines, the performances DO NOT DEGRADE in the name of security. I want perfs, i need perfs. Okay, i spent a lot of money for good and fast disks, i know that loosing a disk on a CCD will require restoring 18GB and that it will takes me 2 or 3 hours but i prefer this eventuality rather than the certainty of 20% performances decrease. whatever you decide in the future, *PLEASE* make it optional. TIA. > In many repsects, write caching could be obsolete soon... ;-). > Not sure IMHO. Typical wait time before write starts is 20 to 50 ms or half the internal disk buffers. Far less than sync daemon. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message