Date: Sat, 23 Dec 2000 23:15:13 -0800 From: "Jake Khuon" <khuon@GBLX.Net> To: Bruce Evans <bde@zeta.org.au> Cc: Jake Khuon <khuon@GBLX.Net>, freebsd-hardware@FreeBSD.ORG, sos@FreeBSD.ORG Subject: Re: fixing hard read errors Message-ID: <200012240715.eBO7FDA10056@Espresso.NEEBU.Net> In-Reply-To: Bruce Evans's message of Sun, 24 Dec 2000 16:42:42 %2B1100. <Pine.BSF.4.21.0012241625010.8602-100000@besplex.bde.org>
next in thread | previous in thread | raw e-mail | index | archive | help
### On Sun, 24 Dec 2000 16:42:42 +1100 (EST), Bruce Evans <bde@zeta.org.au> ### casually decided to expound upon Jake Khuon <khuon@GBLX.Net> the ### following thoughts about "Re: fixing hard read errors": BE> On Thu, 21 Dec 2000, Jake Khuon wrote: BE> BE> > I've started taking corruption errors on the drive. BE> > BE> > Dec 20 23:37:50 Mocha /kernel: ad0: HARD READ ERROR blk# 19527354 BE> > Dec 20 23:37:56 Mocha /kernel: ad0: HARD READ ERROR blk# 19527354 status=59 BE> > error=40 BE> > Dec 20 23:37:57 Mocha /kernel: ad0: DMA problem fallback to PIO mode BE> > BE> > Right now, I've had to set the fsck flag to 0 in /etc/fs and am doing a BE> > forcemount of /usr in /etc/rc before the other mounts just to keep me going. BE> > BE> > Is there a way to use badsect to map these out? When I try, I get this: BE> > BE> > badsect /usr/BAD 19527354 BE> > block 19527354 out of range of file system BE> > Don't forget to run ``fsck /dev/rad0s2f'' BE> BE> The block number printed by the ad driver is relative to the start of the BE> disk, but the block number for badsect must be relative to the start of BE> the filesystem (ad0s2f). So should I just subtract the offset for ad0s2f from 19527354 or do I have to convert the offset to something first based on the blocksize of the file system? BE> Disk drivers should use the standard function for reporting disk errors BE> (diskerr()) so that more information, including the filesystem-relative BE> block number, is printed. From subr_disklabel.c: BE> --- BE> /* BE> * Disk error is the preface to plaintive error messages BE> * about failing disk transfers. It prints messages of the form BE> BE> hp0g: hard error reading fsbn 12345 of 12344-12347 (hp0 bn %d cn %d tn %d sn %d) Yeah... wouldn't that be nice. |8^) -- /*====================[ Jake Khuon <khuon@GBLX.Net> ]======================+ | Senior Manager and Chief Architect, Net-Eng/NSM /~_ |_ () |3 /-\ |_ | | VOX: +1 (408) 543-4828 Fax: +1 (408) 543-0074 \_| C R O S S I N G | +===============[ 960 Hamlin Court, Sunnyvale, CA 94089 ]===============*/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hardware" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200012240715.eBO7FDA10056>