From: tgc@world.std.com (Terry Carlin) Subject: Re: Bad blocks (?) on hard disk Date: Wed, 27 Jan 1993 13:27:20 GMT
ckurs-2@wilbur.uni-mainz.de (Teilnehmer am C-Kurs WS 1992_93) writes:
: In article <1993Jan26.200529.20973@newshub.ccs.yorku.ca>, marcf@nexus.yorku.ca (Marc G Fournier) writes:
: |> In article <1993Jan26.190254.3059@sarah.albany.edu> rob@dinner.asrc.albany.edu (Robert D. Seals) writes:
: |> >Hello,
: |> >I just got sls from tsx-11 today, and prepared things and dove
: |> >in to install on my hard disks.
: |> >
: |> >Things seems sensible except that bad blocks on the hard disks
: |> >aren't mapped to a forwarding area. This makes installation nearly
: |> >impossible, since my disks have many bad blocks. Is there
: |> >an 'fsck' on the a1/a2 disks that can locate and deal with bad
: |> >blocks so that I may proceed with installation? If not,
: |> >what am I supposa do?
: |> >
: |> mkfs -c checks and marks bad blocks
: |>
: |> marc
: |>
:
: No! It does not, but it should! Seems to me that there is a mistake in the
: interpretation of the error register of the HD controller. The only secure way to
: handle the bad blocks problem is to feed the bad block list to mk(e)fs. That is
: what my experieces with mk(e)fs are ... As soon as I find the error, i will post
: a fix for it. (But i haven't found it yet :( )
:
: Dominik
:
Well, my experiences tell me it works at least with the 1542
controller. I just got an old 20 meg SCSI drive which used to sit on
a Mac. I partitioned it and did the mkfs -c /dev/sda1 20000. Worked
well. It found and marked ~15 bad blocks. To test it out, I tar'ed
a 60 meg disk over to the 20 meg drive until it complained about
running out of space. I then did a dd if=/dev/sda1 of=/dev/null and
it went without error. I guess I could really put it through a long
repeating test, but I probably won't put anything on it that would
be critical, 'cause the disk is ~5 years old and may bite the dust at
any sec.
-- Terry Carlin tgc@world.std.com