From owner-freebsd-scsi@FreeBSD.ORG Tue Jan 20 13:33:34 2004 Return-Path: Delivered-To: freebsd-scsi@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 8C4E216A55D for ; Tue, 20 Jan 2004 13:33:34 -0800 (PST) Received: from panzer.kdm.org (panzer.kdm.org [216.160.178.169]) by mx1.FreeBSD.org (Postfix) with ESMTP id 5864443D2D for ; Tue, 20 Jan 2004 13:33:32 -0800 (PST) (envelope-from ken@panzer.kdm.org) Received: from panzer.kdm.org (localhost [127.0.0.1]) by panzer.kdm.org (8.12.9/8.12.5) with ESMTP id i0KLXV64016018; Tue, 20 Jan 2004 14:33:31 -0700 (MST) (envelope-from ken@panzer.kdm.org) Received: (from ken@localhost) by panzer.kdm.org (8.12.9/8.12.5/Submit) id i0KLXVea016017; Tue, 20 Jan 2004 14:33:31 -0700 (MST) (envelope-from ken) Date: Tue, 20 Jan 2004 14:33:31 -0700 From: "Kenneth D. Merry" To: Julian Elischer Message-ID: <20040120213330.GA15900@panzer.kdm.org> References: <400D9B08.70005@vicor.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <400D9B08.70005@vicor.com> User-Agent: Mutt/1.4.1i cc: SCSI@freebsd.org Subject: Re: CAM//SCSI disk timeouts X-BeenThere: freebsd-scsi@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: SCSI subsystem List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 20 Jan 2004 21:33:34 -0000 On Tue, Jan 20, 2004 at 13:18:00 -0800, Julian Elischer wrote: > > > If we are talkign to a raid array and it discovers that it has to > do some maintanance (e.g. declare a physical disk bad or recreate > some data on the fly) it is possible that transactions to the raid > that are outstanding may tak a lot longer than usual. In particular, > we've seen the case where a log entry in teh Raid controller's log > that indicates that a drive has been declared bad, is often accompanied > in FreeBSD with a set of CAM /scsi-disk timeouts > and things degrade from there... > > Where can I find the timeouts in force fo rscsi disks and teh retry > policies? > > I need to compare the theoretical behaviour of FreeBSD CAM with the > theoretical timeouts needed for the raid array, given a disk failure.. You can tweak the timeout and retry count values with the kern.cam.da.retry_count and kern.cam.da.default_timeout sysctl variables. They are also available as loader tunables. Ken -- Kenneth Merry ken@kdm.org