From owner-freebsd-scsi@FreeBSD.ORG Mon Apr 12 17:06:24 2010 Return-Path: Delivered-To: freebsd-scsi@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id CB94D1065676 for ; Mon, 12 Apr 2010 17:06:24 +0000 (UTC) (envelope-from mj@feral.com) Received: from ns1.feral.com (ns1.feral.com [192.67.166.1]) by mx1.freebsd.org (Postfix) with ESMTP id 992AF8FC1E for ; Mon, 12 Apr 2010 17:06:24 +0000 (UTC) Received: from [192.168.221.2] (remotevpn [192.168.221.2]) by ns1.feral.com (8.14.3/8.14.3) with ESMTP id o3CH6Hfl057380 (version=TLSv1/SSLv3 cipher=DHE-RSA-CAMELLIA256-SHA bits=256 verify=NO) for ; Mon, 12 Apr 2010 10:06:20 -0700 (PDT) (envelope-from mj@feral.com) Message-ID: <4BC35309.1010807@feral.com> Date: Mon, 12 Apr 2010 10:06:17 -0700 From: Matthew Jacob Organization: Feral Software User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.9) Gecko/20100330 Fedora/3.0.4-1.fc11 Thunderbird/3.0.4 MIME-Version: 1.0 To: freebsd-scsi@freebsd.org References: <223938752.20100412130323@connection.ca> In-Reply-To: <223938752.20100412130323@connection.ca> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Greylist: Sender DNS name whitelisted, not delayed by milter-greylist-4.2.3 (ns1.feral.com [192.168.221.1]); Mon, 12 Apr 2010 10:06:20 -0700 (PDT) Subject: Re: isp driver - sysctl controls missing X-BeenThere: freebsd-scsi@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list Reply-To: mj@feral.com List-Id: SCSI subsystem List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 12 Apr 2010 17:06:24 -0000 On 04/12/2010 10:03 AM, Ross wrote: > It looks like in the isp(4) driver update that happened back when Freebsd > v8.0 got updated, it looks like the user-modifiable sysctls got > removed and timeouts are now hardcoded. > > The man page for isp state: > -= > SYSCTL OPTIONS > dev.isp.N.loop_down_limit > This value says how long to wait in seconds after loop has gone > down before giving up and expiring all of the devices that were > visible. The default is 300 seconds (5 minutes). A seperate > (nonadjustable) timeout is used when booting to not stop booting > on lack of FC connectivity. > > dev.isp.N.gone_device_time > This value says how long to wait for devices to reappear if they > (temporarily) disappear due to loop or fabric events. While this > timeout is running, I/O to those devices will simply be held. > > -= > > But checking over the driver code shows sysctl details are now removed > and the hardcoded value for loop_down_limit is now 60 seconds and > gone_device_time is 30 seconds. > > Anyone know if there's any reason for the sysctl is being pulled out, > and the limits being so high? We use this in conjunction with geom > multipath, so would prefer to fail over to another fc path asap when > there's an issue. > > > Cheers, > Ross. > > Sounds like a bug that I need to fix.