From owner-freebsd-current Fri Aug 23 17:24:55 1996 Return-Path: owner-current Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id RAA29483 for current-outgoing; Fri, 23 Aug 1996 17:24:55 -0700 (PDT) Received: from GndRsh.aac.dev.com (GndRsh.aac.dev.com [198.145.92.241]) by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id RAA29474 for ; Fri, 23 Aug 1996 17:24:42 -0700 (PDT) Received: (from rgrimes@localhost) by GndRsh.aac.dev.com (8.6.12/8.6.12) id RAA15876; Fri, 23 Aug 1996 17:23:57 -0700 From: "Rodney W. Grimes" Message-Id: <199608240023.RAA15876@GndRsh.aac.dev.com> Subject: Re: -current kills harddrives To: jkh@time.cdrom.com (Jordan K. Hubbard) Date: Fri, 23 Aug 1996 17:23:57 -0700 (PDT) Cc: julian@whistle.com, nirva@ishiboo.com, freebsd-current@FreeBSD.org In-Reply-To: <2790.840838118@time.cdrom.com> from "Jordan K. Hubbard" at "Aug 23, 96 03:08:38 pm" X-Mailer: ELM [version 2.4ME+ PL11 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-current@FreeBSD.org X-Loop: FreeBSD.org Precedence: bulk > > > I seriously doubt that -current is killing your hard drives. > > > Some things you just can't do from software, even if you wanted > > > to. > > > > That's not true. > > > > when we enabled tagged queueing on teh wide busses we > > effectively increased teh work the drive is doing, and we probably > > raised the temperature by another 2 or 3 degrees. If the cooling was in adaquate or right on the edge before, making the drive work harder and producing this extra 2 or 3 degrees of heat could raise the drive temperature into the range that it's life expectancy approches 0. (Micropolis is now shipping drives with temperature tatletell(SP) devices on them with a 71 degree C trip point.) > So we talk to the drives a little faster - you're saying that even > while remaining within spec, simply making the drives work to > performance levels is enough to expect failure? Hmmmm. An > interesting point of view. Yes, you can make hardware break, if you use it in ways that the designer had not atticipated in the design. Case in point, an IBM 1403 line printer (chain style) can be made to physically break the printer chain by a magical sequence of characters that has to be computed for each line of output (You have to know lots of details about the character sequence on the loaded chain, and the timing of the 1403 print hammer bank) with about 120 lines of output. This is caused by the fact that IBM had never intended you to fire 1/3 of the hammers every cycle line after line. Later designs of this same printer added some smarts in the electronics to recongnize the condition of firing more than 1/4 of the hammers in a cycle and added stall cycles to prevent this. Data General 6122 10 platter disk drives could be tossed way out of head alignment by doing full stroke seeks for an hour or two. Some times it would even crash the heads... or overheat the linear motor... -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Reliable computers for FreeBSD