From owner-freebsd-scsi Thu Oct 22 20:40:26 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id UAA15281 for freebsd-scsi-outgoing; Thu, 22 Oct 1998 20:40:26 -0700 (PDT) (envelope-from owner-freebsd-scsi@FreeBSD.ORG) Received: from pluto.plutotech.com (mail.plutotech.com [206.168.67.137]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id UAA15276; Thu, 22 Oct 1998 20:40:25 -0700 (PDT) (envelope-from gibbs@plutotech.com) Received: from narnia.plutotech.com (narnia.plutotech.com [206.168.67.130]) by pluto.plutotech.com (8.8.7/8.8.5) with ESMTP id VAA25994; Thu, 22 Oct 1998 21:39:53 -0600 (MDT) Message-Id: <199810230339.VAA25994@pluto.plutotech.com> X-Mailer: exmh version 2.0.2 2/24/98 To: Don Lewis cc: freebsd-fs@FreeBSD.ORG, freebsd-scsi@FreeBSD.ORG Subject: Re: filesystem safety and SCSI disk write caching In-reply-to: Your message of "Thu, 22 Oct 1998 17:13:09 PDT." <199810230013.RAA19305@salsa.gv.tsc.tdk.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Thu, 22 Oct 1998 21:33:03 -0600 From: "Justin T. Gibbs" Sender: owner-freebsd-scsi@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org >} I've already given my opinion on this. I believe the Hawk is seeing >} a power glitch or temporary power loss when the reset switch is hit and >} so the contents of the cache are lost. I have never said that the >} behavior that Don Lewis is seeing is 'not possible', only that, for >} the drive in question, the reset causing cache corruption is not likely. > >If this problem caused by a load related power glitch, then it is >possible to get silent filesystem corruption in normal operation if >write caching is enabled, since cached writes could get lost and the >driver might never notice. The driver will notice as the drive will notice an issue a Unit Attention response the next time you touch it. Or is your point that you won't necessarily access the device again and so never see that the device saw a loss in power? There has been quite a bit of debate on how UAs should be handled. The original CAM driver was *very* conservative and returned all pending I/O with EIO, marked the pack invalid, and refused to take any I/O unless the device cycled through final close. This ensures that if the device or pack is replaced that you don't spam different media or even if the media is the same, make the problem worse by attempting to continue after some transactions were irrevocably lost. This was considered too disruptive and since a permanent solution could not be developed for 3.0R, the UA code was disabled. The correct solution likely requires better communication to the FS or user layer so that pack validation of some sort can occur. -- Justin To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-scsi" in the body of the message