From: becker@super.org (Donald J. Becker) Subject: Re: HD Bad Sectors Date: Tue, 16 Feb 1993 02:00:30 GMT
In article <C2I2G5.M8H@nmrdc1.nmrdc.nnmc.navy.mil> dsc3pzp@nmrdc1.nmrdc.nnmc.navy.mil (Philip Perucci) writes:
>In article <40947@sophia.inria.fr> vasselle@aragorn.inria.fr (Bruno Vasselle) writes:
>>
>> I've got troubles with sectors I've marked "bad" when I formated
>>my hard disk at a low-level. It seems that linux tries to read/write on them,
>>even if it recognizes them as bad sectors (error 0x80). I think someone has
>>remarked the problem in this newsgroup a few days ago, but I didn't see any
>>response.
>>...
>For poor-folk like us with MFM/RLL/ESDI drives, this is DEFINATELY an issue.
>The trick is to first produce a list of bad blocks in a form suitable
>for "mkfs -l". "mkfs -c" won't do it for us. I am going to try and
>write a utility to automate this. Maybe...
I think writing such a utility is a poor solution. I can see the
advantage in something like:
collect_bad_blocks /dev/hda2 | mk*fs -l - /dev/hda2
but if the current mechanism is a good one (and in this case it seem
to be somewhat better), you should fix that instead.
I recently ran into this problem (mk*fs -c failing to collect the bad
blocks using the 0.99.4 kernel) myself. The solution was upgrading to a
0.99.5 kernel. While this may be difficult for people trying to use
the current SLS distribution (which uses 0.99.4), the solution is to
wait for the fix to filter through the "system", or find someone that
will make a new a1 disk for you.
Making a new 'a1' disk is easy if you can already make a kernel. Set
ROOTDEV=/dev/fd0 and RAMDISK=-DRAMDISK=688 in the top-level Makefile,
and copy the resulting kernel onto a disk already written with SLS/a1.
-- Donald Becker becker@super.org Supercomputing Research Center 17100 Science Drive, Bowie MD 21114 301-805-7482