From owner-freebsd-qa Sat Dec 29 23:11:45 2001 Delivered-To: freebsd-qa@freebsd.org Received: from hotmail.com (f204.law9.hotmail.com [64.4.9.204]) by hub.freebsd.org (Postfix) with ESMTP id A5F8B37B41A; Sat, 29 Dec 2001 23:11:38 -0800 (PST) Received: from mail pickup service by hotmail.com with Microsoft SMTPSVC; Sat, 29 Dec 2001 23:11:38 -0800 Received: from 24.9.137.53 by lw9fd.law9.hotmail.msn.com with HTTP; Sun, 30 Dec 2001 07:11:38 GMT X-Originating-IP: [24.9.137.53] From: "Chuck T." To: dwhite@resnet.uoregon.edu Cc: qa@FreeBSD.ORG, stable@FreeBSD.ORG Subject: Re: 4.5-prerelease (27 dec) - UDMA33 ICRC error trashed fs Date: Sat, 29 Dec 2001 23:11:38 -0800 Mime-Version: 1.0 Content-Type: text/plain; format=flowed Message-ID: X-OriginalArrivalTime: 30 Dec 2001 07:11:38.0629 (UTC) FILETIME=[395A2B50:01C19101] Sender: owner-freebsd-qa@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG >On Sat, 29 Dec 2001, Chuck T. wrote: > > > I seem to have trashed my root file system on a 4.5-prerelease system > > cvsup'ed on 12/27. I'd installed 4.5-prerelease on a virgin (master) >drive, > > and had been copying some configuration files from the 4.4-release >(slave) > > drive. When I was copying the files I noticed a couple of UDMA ICRC >errors > > before the kernel it fell back to PIO. I didn't notice any further >problems > > until I moved the 4.4-release drive back to the master and tried to boot > > from it. The drive would not boot at all. I again reconfigured to boot >from > > the 4.5 drive with the (thrashed) 4.4 drive as slave. Running fsck >reported > > tons of errors (a log is available if needed). > >I seriously doubt this is a FreeBSD issue. Ok. > >Here's the relevant messages: > > > Dec 28 09:03:09 fastbsd /kernel: pcib0: PCI > > bridge> on motherboard > >First, 440BX chipset, pretty standard fare. I've run hundreds of these >boxes with IDE without problems. > > > Dec 28 09:03:09 fastbsd /kernel: ad0: 38166MB > > [77545/16/63] at ata0-master UDMA33 > > Dec 28 09:03:09 fastbsd /kernel: ad1: 6149MB >[13328/15/63] at > > ata0-slave UDMA33 > > Dec 28 09:03:09 fastbsd /kernel: acd0: CDROM at > > ata1-master using PIO4 > >Western Digital slaving a Western Digital. Plus a CDROM on the >other channel. Yup. Anything known to be bad about a Western Digital slaving a Western Digital ? > > Dec 28 09:22:11 fastbsd /kernel: ad1a: UDMA ICRC error reading fsbn >393471 > > of 196704-196719 (ad1 bn 393471; cn 416 tn 5 sn 36) retrying > > Dec 28 09:22:11 fastbsd /kernel: ad1a: hard error reading fsbn 393471 of > > 196704-196719 (ad1 bn 393471; cn 416 tn 5 sn 36)ad1: DMA problem >fallback to > > PIO mode > >'hard error' is generally Bad(tm). If it's been a while since you've power >cycled this drive, it looks like it just went south on you. I would >suggest downloading the Diagnostic toolset from Western Digital's site and >scanning the drive for problems, and fixing them if possible. The drive had only been powered up for a few minutes. Note that once the kernel fell back to PIO mode everything was fine. I only copied a few K (my hosts file, inetd.conf, rc.conf, and XFree86.conf). I recovered as much of the root partition as I could (in pio mode) and the rebooted from it and ran for several hours without further incident w/ 4.4 (in a single disk configuration). The /usr and /var partitions were both fine. I think my problem is excessive cable length plus drive drawers ... as I said in my original post I seldom run with two drives under FreeBSD. I reported the problem since it appeared to be a new issue with 4.5-prerelease. In retrospect I should have mounted the drive as read-only since I was only going to read from it! I was quite surprised that it got trashed (aren't we always!). _________________________________________________________________ MSN Photos is the easiest way to share and print your photos: http://photos.msn.com/support/worldwide.aspx To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-qa" in the body of the message