How to map a bad block number to a file & vice versa

How to map a bad block number to a file & vice versa

Post by James Lon » Wed, 04 Jul 2001 14:16:50



My security report is giving me a block number.
How can I find out what file is stored there?
Is there a better way than to cp -R / /dev/null
and see which file craps out?

Also, is there a way that I can create a file
and force it to occupy one block (512 bytes?)
and specifically, that particular block?

Anybody here remember UCSD Pascal?  Back in the
8" floppy days, we just wrote a file BADSPOT.BAD
to cover the bad block and went on.  At this point,
it is the only bad block I see, the system is rela-
tively static as far as configuration changes or
accumulated data, and I'm backing up weekly, so I'm
willing to write a badspot file over that spot and
live with it.

Thanks in advance.

 
 
 

How to map a bad block number to a file & vice versa

Post by Steven G. Kar » Wed, 04 Jul 2001 14:46:02




Quote:> My security report is giving me a block number.
> How can I find out what file is stored there?
> Is there a better way than to cp -R / /dev/null
> and see which file craps out?

> Also, is there a way that I can create a file
> and force it to occupy one block (512 bytes?)
> and specifically, that particular block?

Is this a scsi disk?  Use camcontrol to enable auto
read/write reallocation for bad blocks.  If this an eide
disk, I suggest replacing the drive.

--
steve

 
 
 

1. Using Oracle to mirror NIS maps & vice-versa

   Has anyone ever used Oracle to "mirror" NIS maps, as well
as conversely?  I work at a company as UNIX programmer and systems
administrator in my group and we manage close to a thousand
workstations (Suns, SGIs, Solbournes) and a few dozen servers.
We also have 2 servers implemented as NIS masters (hence, two
different domains).  We would like to implement Oracle in
such a way that the database has tables which essentially
mirror the NIS maps (one table per map, say?).  The scheme is that
Whenever a change to a map occurs (via /var/yp/make map_name) the
corresponding table in Oracle gets changed as well.  A more
difficult scheme would be to do the reverse (update the database
table and the corresponding NIS map would automatically be
updated), but my main concern is getting the former implemented
first.  
   Has anyone ever done this before?  How feasible is it?  What
should I be concerned with in such a scheme?  BTW, we also plan
to create tables which do not have corresp. NIS maps but will
be used to hold other useful info.  The purpose for thsi whole
scheme is to enable users and administrators to tap into the
Oracle database as a global source of information on our workstations
(as well as PCs ... soon), as well as users of thos workstations,
instead of relying on several different sources for the information
(e.g. WingZ spreadsheets, FrameMaker documents, files out in UNIX,
as well as NIS maps).  One could easily tap into the database via
shell scripts or a variety of programming languages.

    Any opinions or help on this matter is greatly appreciated!

                                --John

2. help need driver for our T1 card

3. Can I take files from dos to unix & vice versa?

4. How to let key arrow retrive history commands when remote login sh

5. Trying to MAP a drive from AIX to NT 4.0 or vice versa!

6. Cursor repeats operation

7. Mounting Solaris disks on Windows 95 & vice versa....

8. Linux on Compaq Deskpro 2000 MT 5133

9. how to backup files from solaris 8 to linux rh 6.0 or vice-versa at SOHO.

10. copying files from my directory to a floppy disk or vice versa

11. Can you mount Linux file systems and vice versa

12. Transfering file from server to cliant and vice versa

13. NT can see Linux & vice versa