[SUMMARY] disk problem

From: Yann JOBIC <jobic_at_polytech.univ-mrs.fr>
Date: Tue Nov 07 2006 - 12:19:23 EST
Thanks to the following Experts :
john stoffel, Dana Sparling, Dana Hudes, Hal Huntley, Gregory Shaw, 
Thomas Payerle

The usual method for saving a disque is to use the format utility, with 
the read analyze command.
given by Thomas)
I tried that but that just don't work. The raison was pointed by john :
I've got an IDE drive and we can't re-map the bad blocks.

John gave me another good advice : there's a little soft, called 
dd_rescue, that can save most part of the hard drive.
Briefly, dd_rescue is like plain old *nix dd, except that it acts 
intelligently when it finds bad sectors. It reads the source with a big 
block size (fast!) while everything is OK and, when it hits an error, 
falls back to a small block size (slow) to try to get everything except 
for the bad block.
(source : http://blogs.sun.com/superpat/category/General)

Actually i didn't tried that yet.

I could mount the disk in read mode :
mount -o ro /dev/dsk/c0t0d0s0 /mnt
And grabe the license files that i needed. I've got now a running 
license server, and i've got more time to play with dd_rescue.

Original question :
> Hi,
> I've got a corrupt filesystem on an ultra 5 machine running solaris 8.
> I've got some bad blocks. I know them, fsck told me, but it won't tag 
> them as dead.
> Running fsck 2 or 3 times also doesn't help. ( the blocks are : 907699 
> to 907711 )
> I tried to run fsck on some different superblocks, as below :
> fcsk -o -b=n -y /dev/rdsk/c0t0d0s0 ( n given by the command : newfs -N 
> /dev/rdsk/c0t0d0s0 )
> That doesn't help.
> I'm running out of ideas. How can i tag the bad blocks as bad blocks ?
> I would like to dump the disk, but i can't so far.
> Thanks in advance,
> Yann


HPC engineer
Polytech Marseille DME
TechnopC4le de ChC"teau Gombert
5 rue Enrico Fermi
13453 Marseille cedex 13
Tel  : (33) 4 91 10 69 37
Fax : (33) 4 91 10 69 69 
sunmanagers mailing list
Received on Tue Nov 7 12:18:21 2006

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:44:02 EST