Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!usc!sdd.hp.com!hplabs!otter.hpl.hp.com!otter!nt
From: nt@otter.hpl.hp.com (Nicolas Tripon)
Newsgroups: comp.os.minix
Subject: marking bad blocks
Message-ID: <78380002@otter.hpl.hp.com>
Date: 30 May 91 10:58:41 GMT
Organization: Hewlett-Packard Laboratories, Bristol, UK.
Lines: 30


  On a 386sx AT compatible, with a ST506 controller and a Seagate-125 20 Mo
hard disk, I made three partitions : 1 of 1 Mo for DOS, 2 and 3, of 17 Mo and
2 Mo, for Minix 1.5. Then :
# readall -b /dev/hd2 >bad.2    - 18 bad blocks are reported
# sh <bad.2                     - there are 3 .Bad_0x files, summing up to 18 Ko

# readall -b /dev/hd3 >bad.3    - 27 bad blocks reported
# sh <bad.3
Bad block number <x> : this block is marked in bitmap.
  "
Bad block number <x+8> : this block is marked in bitmap.
Temporary file /tmp/? already in use.

Mounting /dev/hd3 and listing the files, I found 4 files .Bad_000x of length 0.
Restarting the whole operation and rebooting between 'sh <bad.2' and
'readall -b /dev/hd3 >bad.3', 'sh <bad.3' does not complain any more, and the
file system shows 27 Ko in 4 .Bad_0x files. But, after copying files in 
/dev/hd3, at reboot, I get :

Unrecoverable disk error on device 3/3, block 891
   "                                          893
Unrecoverable disk error on device 3/3, block 899
Unrecoverable disk error on device 3/3, block 900

Blocks 891 to 899 were reported by readall and supposed to by out of the way 
of the controller ; block 900 and following were not detected as bad by 
readall. Please note that I use the Universal boot diskette, as the AT boot
triggers a kernel panic in readall, after meeting several bad blocks.
  Can anyone help ? -Nicolas Tripon