fsck fails - mark sectors as bad?

[Dan Langille]

On 29 Mar 2004 at 14:29, Bjoern Fischer wrote:

> Hello Dan,
> 
> > I had a hard drive failure.  I'm trying to get as much data off it as
> > I can before I restore from backup.  I have mounted the drive in
> > another box and I'm attempting to salvage what I can.
> 
> being confronted with disk crashes like the one you described I usually
> try to make a last backup of the complete unmounted file systems with dd.
> Trying to fix the file system in place, i.e. on the failing disk that is
> about to break down completely, makes the situation worse in many cases.
> After dd'ing the file systems to a safe disk you can try fsck.

I'm finding the dd is challenging:

# dd if=/dev/ad2s1e of=ad1s1e conv=noerror,sync bs=64k
dd: /dev/ad2s1e: Input/output error
6+0 records in
6+0 records out
393216 bytes transferred in 42.776918 secs (9192 bytes/sec)
dd: /dev/ad2s1e: Input/output error

- note, this dd has not finished, I think because of these errors 
from /var/log/messages

Mar 28 19:47:53 tmp /kernel: ad2: READ command timeout tag=0 serv=0 - 
resetting
Mar 28 19:47:53 tmp /kernel: ata1: resetting devices .. done
Mar 28 19:48:03 tmp /kernel: ad2: READ command timeout tag=0 serv=0 - 
resetting
Mar 28 19:48:03 tmp /kernel: ata1: resetting devices .. done
Mar 28 19:48:13 tmp /kernel: ad2: READ command timeout tag=0 serv=0 - 
resetting
Mar 28 19:48:13 tmp /kernel: ata1: resetting devices .. done
Mar 28 19:48:21 tmp /kernel: ad2s1e: hard error reading fsbn 1279871 
of 768-895 (ad2s1 bn 1279871; cn 79 tn 170 sn 26) trying PIO mode
Mar 28 19:48:21 tmp /kernel: ad2: DMA problem fallback to PIO mode
Mar 28 19:48:21 tmp /kernel: ad2: DMA problem fallback to PIO mode
Mar 28 19:48:26 tmp /kernel: ad2s1e: hard error reading fsbn 1279903 
of 768-895 (ad2s1 bn 1279903; cn 79 tn 170 sn 58) status=59 error=40
Mar 28 19:48:40 tmp /kernel: ad2: READ command timeout tag=0 serv=0 - 
resetting
Mar 28 19:48:40 tmp /kernel: ata1: resetting devices .. ata1-slave: 
ATA identify retries exceeded
Mar 28 19:48:40 tmp /kernel: done
-- 
Dan Langille : http://www.langille.org/
BSDCan - http://www.bsdcan.org/