questions about bad blocks

Sun Mar 5 17:35:58 AEST 1989

In article <732 at bagend.UUCP> jan at bagend.UUCP (Jan Isley) writes:
|>I was formatting a disk Monday and when it finished formatting, it 
|>announced that the drive had 14 entries in the bad block table.
|>This was *before* the media test was run and I did not enter any
|>bad blocks before I formatted the drive ... fresh out of the box.
|>
|>What gives?  How does it know that there are bad blocks before the
|>surface test is run?
|>
Yes, according to most disk drive manufacturers (and something Gil 
told me) the bad block table is in a standard spot on the hard disk, and
is in a standard format.  So when the manufacturer puts the drive
through bad block checking, it will write a bad block table.  I was
doing some *low level* formatting of a CRC Wren II the other day on a
3B2, and I was playing with the drive on a 3B1, it knew the bad blocks
that we added with the 3B1 diagnostic disk, on the 3B2.  

|>Exactly where is the bad block table and how do I edit/delete it?
|>
You can edit it with the "Enter bad block option" from the diagnostic
disk.  At least you can add blocks, I don't know if you can delete blocks.
My suggestion if you want to remove all the blocks in the bad block table
is to zero it.  Make sure you have a good list of bad blocks with the
iv -vt /dev/rfp000 (or appropriate drive if you are fortunate to have
two drives ;-) )  before you start zeroing bad blocks.

Here's part of something I posted a while back that same subject ...

...
Here is a method used to CLEAR the entire bad block table.  Be careful,
this is dangerous.  Make sure you perform a surface test afterwards.  It
would be a good idea to enter the bad blocks that was found on the 
drive by the manufacturer (usually on the label on the top of the drive,
again open the case -- don't be afraid)

Load diagnostics.
type:	s4test
expert> 31
i> i			      <-  Initialize
i> dr 1                       <-  Hard Disk 1
i> i                          <-  Initialize
i> sb 0                       <-  Set sector buffer to 0's
i> wr 2                       <-  Write to sector 2
i> rd 2			      <-  Read sector 2 into buffer
i> sb                         <-  Display buffer
i> Q
expert> U

|>How can I add drives to the list on the diagnostics disk?
|>
|>Lenny, your list is different than the 3.51 list ... anyone got 
|>the source code?
|>
No, I don't have the source code.  That diagnostic disk which has the
*bigger* drives like the Maxtor XT1140 and XT2190 in there was something
that someone with the sources did.  You cannot add drives to that list
without the source, and I don't have it :-(  You can, though, use the
option for "Others" and answer the questions appropriately for the
drive parameters to format any kind of disk drive.  Therefore, you don't
need a specific option for your kind of drive.

|>After I ran the surface test and loaded the foundation set, I ran
|>
|>	iv -s /dev/rfp000
|>
|>I have never had a problem doing this before, but this made the 
|>system crash every time.
|>
This can be dangerous.  Doing a surface check on a mounted filesystem
that is in use ... you are playing with the superblock, the vtoc, and
the bad block table on a running system.   If you were going to do this by
hand I would suggest booting the floppy UNIX and then umount'ing the
/dev/fp002 first.

...
|>When you get something like this in /usr/adm/unix.log:
|>
|>	drv:0 part:2 blk:43763 rpts:1 Tue Jan 24 16:15:32 1989
|>	drv:0 part:2 blk:43763 rpts:1 Tue Jan 24 16:35:02 1989
|>	drv:0 part:2 blk:43763 rpts:2 Tue Jan 24 16:38:47 1989
|>	drv:0 part:2 blk:43763 rpts:1 Tue Jan 24 16:43:12 1989
|>
|>Is this a bad block looking for a place to happen?  Yea, I know you can
|>use the recently posted blockfind to find the offending file.  But, do
|>you panic immediately, find the file, copy it, add the bad block to the
|>table, delete the old file and fsck the disk, .... or what?
|>Do you do this on the first offense you notice?  Or do you wait for
|>*possible* problems later?
|>
Well it wasn't the first offense.  It was over a 1/2 hour period of time and
you had 5 reports of that bad block.  If the data at that block (if it is
allocated to a file) isn't important, I would map the bad block.  

-Lenny

-- 
Lenny Tropiano             ICUS Software Systems         [w] +1 (516) 582-5525
lenny at icus.islp.ny.us      Telex; 154232428 ICUS         [h] +1 (516) 968-8576
{talcott,decuac,boulder,hombre,pacbell,sbcs}!icus!lenny  attmail!icus!lenny
        ICUS Software Systems -- PO Box 1; Islip Terrace, NY  11752