Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 7 Oct 2000 09:58:23 +0200 (CEST)
From:      Remy Nonnenmacher <remy@boostworks.com>
To:        jedgar@fxp.org
Cc:        dest@syd.eastlink.ca, freebsd-stable@FreeBSD.ORG
Subject:   Re: Bad IDE Drive
Message-ID:  <200010070757.JAA39707@luxren2.boostworks.com>
In-Reply-To: <20001006152247.B82507@pawn.primelocation.net>

next in thread | previous in thread | raw e-mail | index | archive | help
On  6 Oct, Chris Faulhaber wrote:
> On Fri, Oct 06, 2000 at 04:14:38PM -0300, Craig Hawco wrote:
>> Hello,
>> 
>> 	I have recently fled from FreeBSD back into (ugh) windows because of a 
>> minor drive problem. It seems that my drive has a few bad blocks, and I 
>> know what they are. FreeBSD seems to try to write to the same bad sectors 
>> every time, and keeps printing errors that it can't write to the block, etc 
>> etc. The thing is it's the same few (about a dozen) sectors, about 8/10ths 
>> the way through the drive. I could use only up to that amount, but that 
>> doesn't seem like an elegant solution leaving a few hundred megs of empty 
>> space. Is there a way to have FreeBSD map the sectors and try to neither 
>> read nor write to them? It wouldn't be a problem (I can turn off syslog and 
>> get rid of the nasty messages ;) if it didn't cause the filesystem to 
>> become corrupted (fsck tries to read/fix the data on the bad spots and 
>> dies). Hope someone out there can help.
>> 
> 
> Most modern IDE drives automatically remap bad sectors.  If the drive can
> no longer remap bad sectors itself, it is probably time to replace it.
> 

Remapping works for IDE and SCSI but most of them remap only on a write
operation, not on a read (Even is told to do so by ARRE for SCSI disks).
This is intended to let you know that something is wrong, instead of
silently reallocating a new zeroed sector and let you then, later,
guess why your data are corrupted.

Unfortunetly, this require you to overwrite the sector (and more
generally the 8K block covering the sector(s), else the system will try
to read in the block in buffers before writting and will fail the
write).

To get things worst, some disks succeed in rewritting and do not really
relocate the sector, making a Damocles' sword of the problem. Thanks to
camcontrol (or scsi(1)), you can force rellocation on SCSI disks pretty
easily.

RN.
IeM





To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200010070757.JAA39707>