> Unfortunatly, I didn't make the first installation on this server, so I
> don't know exactly the drive configuration. It have 2220 sectors, it
> seems that the partition drive looked like
> 1 swap drive
> 2 /
> 3 /usr
> 4 /boot
> I think we almost found /usr wth scanext2. It's a bit strange, the
> partition can be mounted, i can do ls but on directories I got I/O
> error. What could I do to fix that ? Can this be because the end of the
> partition is too close from start, too far ? or a probleme whith head or
> cylinders in partition définition ?
> Wolfgang Fischer wrote:
>> first of all, you should make a backup from your whole disk on another
>> dd if=/dev/hdx (or sdx for SCSI drives) of=/path/to/backup bs=10M
>> Then you can try out some tools without destroying data.
>> If you know exactly the size of your partitions, you can restore the
>> partition table using (c)fdisk.
>> On Sun, 13 Jul 2003 04:00:11 +0200, Rodolphe SIBERT wrote:
>> You should really have backups of your production servers.
> We have backup but there are not complete due to recent change in an
> application directory. 75% only of the work is saved...
> I think so. It's a raid 5 with 3 disks, one disk fail (see bleow) but
> the two last are fine
>>>The problem come after a loosy wizard on a bootable cd erase the partition
>>>table and create a 50Mo fat 16 partition (no format have been done).
>> Don't do this on a production server without knowing what you are doing.
> 200 % agree with you. I didn't launch the cd ... I'm just trying to get
> things back ...
> At first, a problem on the ServeRAID controler (one DDD disk) lend us to
> put the IBM hardware setup cd to access the ServeRAID manager (It's only
> available for X and no X on this server nor on any linux where it
> lives). After the raid setup done, the wisard continue on the root
> partition creation which for sure was a *very* bad idea !