From owner-freebsd-stable Fri Apr 12 10:29:31 2002 Delivered-To: freebsd-stable@freebsd.org Received: from mailhost.firstcallgroup.co.uk (dilbert.firstcallgroup.co.uk [194.200.93.142]) by hub.freebsd.org (Postfix) with ESMTP id 6B96137B405 for ; Fri, 12 Apr 2002 10:29:25 -0700 (PDT) Received: from pfrench by mailhost.firstcallgroup.co.uk with local (Exim 3.34 #1) id 16w4rP-000Klk-00 for stable@FreeBSD.ORG; Fri, 12 Apr 2002 18:29:19 +0100 To: stable@FreeBSD.ORG Subject: SMP problems with latest SUP and SCSi drives Message-Id: From: Pete French Date: Fri, 12 Apr 2002 18:29:19 +0100 Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG A few days ago I posted a problem with a SCSi bus on a stable machine allegedly being reset according to the messages from the kernel. I have now had time to track this down and ensure it isnt a hardware fault The symptoms are as follows: copy a set of files (e.g.): cp src/cgi-bin/**.exe copy a set of files (e.g.): cp src/cgi-bin/*.exe /usr/local/www/beta/cgi-bin the reopeat the command to a different directory cp src/cgi-bin/*.exe /usr/local/www/live/cgi-bin and the disc freezes - if left long enough it usually generates an error of the form "sym0:(0,-1,-1) scsi bus reset detected".... hats from memory as I dont have the log to hand. note that the copy involves moving files from ne disc to another - both discs being on the same scsi bus. I have, however, acertained that this *only* happens when the machine is running an SMP kernel. With a non SMP kernel it is fine. With an older version of -STABLE it is also fine (frustratingly I no longer have the date of that older kernel, but we are talking post 4.5_RELEASE). The system is a Compaq Proliant and I am using the onboard SCSI controller. Its has 256 meg of memoiry and a pair of 9.1GB drives attached to sym0. tthe dmesg was posted in the previous email ont eh subject (along with the sym0 timeout message) help! obviously something is going on here under SMP that isnt happening on a single porcessor box. How can I set about generating more information to help track this bug down ? thanks for any assistance, -pcf. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message