I have tried using sdremap to remap what appears to be a faulty block
on a SCSI disk (i386 arch) - am I missing something? 

Testing with;

  % dd if=/dev/sd1h conv=noerror of=/dev/null
  % grep 'sd1(ahc1:1:0): medium error' /var/log/messages
    ...
    /netbsd: sd1(ahc1:1:0): medium error, info = 3980364 (decimal), \
        data = 72 67 00 00 00 00 00 00 00 00
    ... (there are a number of blocks reported)
  % umount /dev/sd1h
  % sdremap /dev/rsd1d 3980364

Retesting then produces the same error.
        
Brian Buhrow writes:
>         I've also found that many of the SCSI disks we use around here report
> that they are having trouble reading bad sectors, but that they do *NOT*
> automatically remap them to new sectors.  Using SunOS's format utility to
> remap them or Jason Torp's sdremap program to remap them under NetBSD has
> proven to be quite successful at breathing new life into disks for several
> years.  If we see  a lot of errors on a disk, we know it's pretty much
> toast.  However, if we get a few persistently bad sectors, a remap usually
> fixes things up quite nicely.


------------------------------------------------------------
Rex McMaster                            rmcm@compsoft.com.au 
                                   rex@mcmaster.wattle.id.au
     PGP Public key: http://www.compsoft.com.au/~rmcm/pgp-pk