BADSPOTS(1M-SysV) RISC/os Reference Manual BADSPOTS(1M-SysV)
NAME
badspots - map out bad spots on hard disks
SYNOPSIS
/etc/badspots [-a -clogfile [-mmaxretries] [-pscanpasses]
[-d] [-e] [-rrange]]
| [-s -vvolume -b(block[,block] | c/h/s[,c/h/s])] -mmaxre-
tries] [-e]] | [-t -vvolume] | [-i]
DESCRIPTION
badspots is used to display and map out bad spots on hard
disks. There are four modes in which badspots can be
invoked. If badspots is invoked without any arguments, -i
is assumed.
-a badspots will scan the console log file specified with
the -clogfile flag. All disk error message must be in
one of the following formats to be accepted:
Since the reported block (sector) number isn't neces-
sarily the one containing the defect, either -rrange
blocks or 16 blocks on each side of the reported blocks
will be scanned for defects. It may take a number of
attempts to find the defects if the error happens
intermittently. badspots will try 3 times unless
-pscanpasses is specified, in which case that number is
used. If -d is specified, the defective sectors will be
displayed, but won't be mapped out. If -e is speci-
fied, the defect(s) will be mapped out, even if an
error occurred that could possible cause data integrity
to be lost. Otherwise, badspots will only map out the
defects(s) if it can verify data integrity. It will
attempt to read the data 5 times unless -mmaxretries is
specified.
-s Map out specific blocks. On SCSI disks, the list of
logical blocks is specified as
-bblocknumber[,blocknumber]... On SMD disks, either a
list of logical blocks can be specified, or the list
can be in in the form of
-bcyl/head/sector[,cyl/head/sector]... If -e is speci-
fied, all of the blocks in the list will be mapped out
even if errors occur while checking for data integrity,
and the number of retries has been exhausted.
-l List the defects on the specified volume. If the
volume is an SMD disk, badspots will list those defects
specified on the bad sector table stored in the volume
header partition of the disk. For SCSI drives, the
primary and growth defects that are stored on the drive
will be listed.
Printed 1/15/91 Page 1
BADSPOTS(1M-SysV) RISC/os Reference Manual BADSPOTS(1M-SysV)
-i Starts an interactive session, beginning with the fol-
lowing prompt:
Enter path to /dev entry for entire volume [q to quit]
:
The full path name of the raw partition containing the
entire volume should be entered. This is usually par-
tition 10.
(l)ist defect table,(a)dd defects,(s)can for
defects,(q)uit [q] :
(l) Lists the defects, similar to the -t mode.
(a) For an SMD disk, badspots will first prompt for
the format in which the defects will be entered
(blockno or c/h/s).
For either SMD or SCSI disks, badspots will then
ask for a list of defects. Only those defects that
had no data errors will be mapped. To map out the
disk even when errors occur, invoke badspots in -s
mode and specify the -e flag.
(s) After prompting for the number of scan passes, the
beginning and ending block numbers, a read-only
scan of the drive is done to check for ECC errors.
Those defects are then displayed.
In addition to the four command line options to specify in
which mode to run badspots, the following is a list of the
additional flags which can also be used in conjuction with
the mode flags.
-clogfile specifies the name of the file containing the con-
sole log that will be scanned by the -a option.
-rrange Since data transfers through the filesystem are
done 8 Kbytes at a time, the block reported as
being in error isn't necessarily the exact block
in which the error is occuring. Usually it is the
first block of the transfer. When the -a option is
used, range sectors on each side of the reported
block are also scanned for errors. If range isn't
specified, a value of 16 is used.
-mmaxretries
If an error occurs when trying to map out the
defects, badspots will retry up to maxretries
times before aborting. The default value is 5.
Page 2 Printed 1/15/91
BADSPOTS(1M-SysV) RISC/os Reference Manual BADSPOTS(1M-SysV)
-d Once bad spots have been found while scanning the
errors reported in the console log, specifying -d
will only display the defects, and not map them
out.
-e Map out bad spots even when errors occur. Nor-
mally, the bad spots aren't mapped out if errors
occur.
-bblockno[,blockno...] or -bc/h/s[,c/h/s...]
List of block numbers to map out under the -s
option.
-vvolume The name of the raw partition which contains the
entire disk. This is usually partition 10.
-pscanpasses
The number of times to scan the range of blocks
found when the -a option is specified. The
default when this isn't specified is 3.
SEE ALSO
format(1M), dvhtool(1M).
NOTES
Defects are only added, and never modified or deleted on
SCSI drives. Therefore, each time a logical block number is
added, it gets reassigned. No checking is done to see if
the block number was previously entered, since that informa-
tion isn't stored anywhere.
For SMD drives, when a logical block number is entered that
is already on the list, it assumes that it is a duplicate.
However, since format deals in logical blocks, it could pos-
sibly happen that the sector immediately before this sector
had previously gone bad, and the entire track needs to be
mapped. If this is the case, run badspots under -s mode, and
map out the block immediately after it.
There also exists a corresponding case for bad sectors
occuring on a track that has already been remapped. format
assumes that the defect is on the logical track that was
specified, when in fact the defect now exists on the track
in the replacement partition to which it was mapped. If
this occurs, dump the bad sector list to determine which
track in the replacement partition is was mapped to, and
then map out a block on the mapped track. Caution must be
taken when doing this, since badspots doesn't correctly
reconfigure the replacment partition to prevent that mapped
track from erroneously being mapped to yet another track in
the replacement partition.
Printed 1/15/91 Page 3