Re: hard disk crash

From: Jose Tabisi (josetabisi_at_tecnitower.com.ar)
Date: 07/25/03 

Date: 25 Jul 2003 07:05:51 -0700

Thanks Steve

Regarding your post scriptum, it looked like a good solution until it
failed. I had bad experiencies with tape backups in the past, of
course now I regret not having other backups around.

Back to your message, I did send the main disk to an expert, there are
not as many choices here as there are there, and I hope to have made
the right choice, I will know soon.

However I have kept the second disk and keep on working on it. I
downloaded a disk editor for windows, mounted the unix disk as
secondary and looked at the info, it seems to be still there.

I was tring to figure what the original partitions were because I
really do not know. Unix data seems to start somewhere around cylinder
750, which makes perfect sense considering /boot must lie within the
first 1024 cylinders of the disk for Unix to be able to start. I then
mounted the disk as secondary with a working Unix as master disk and
tried different fdisk partitions followed by 'divvy /dev/hd10'. So far
I have not been succesfull, maybe there's something wrong with my
reasoning.

If I did try, for example, using the whole disk for the Unix partition
(which was not the original partition), is there any chance I can get
back the data ? Any ideas ? Thanks for your help

Steve Fabac <smfabac@att.net> wrote in message news:<3F1FB8FA.72D65064@att.net>...
> Jose Tabisi wrote:
> >
> > I am working with SCO Unix 5.0.4, using two disks, the second one with
> > a removable bay so that I can use for backup purposes. By mistake we
> > powered on the machine with both disks connected as master disks to
> > the same IDE controller. This usually meant that the machine would not
> > start until that problem got corrected, but this time it somehow
> > managed to start and then finally crashed after a while. Both disks
> > have a DOS and UNIX partition, and the UNIX partition have /root and
> > /u divisions defined.
> >
> > Now the situation is that both disks have lost the partition
> > information and I can not access them in any way. I have searched and
> > downloaded some demo programs available in Internet and I have been
> > able to scan, list and even recover files which reside in the DOS
> > partition. However, none of these programs seem to be able to
> > recognize HTFS files. Both disks seem to be both physically and
> > logicaly intact, the BIOS finds them OK and the scanning of the
> > programs do not end with an error condition.
> >
> > What I do have available are the root and boot disks. What I do not
> > have are either partition information and the divisions info. All I
> > really need is the information which resides in the /u division of the
> > UNIX partition, the rest of the disk I donīt care.
> >
> > Any comments and suggestions will be highly appreciatted.
>
> If the data on these disks is valuable to you, get professional help.
>
> If you have the necessary records then try these steps:
>
> 1) Boot the Emergency boot/root floppies and use fdisk to setup the
> UNIX partition with start block and size identical to the settings used
> for the original install. If you're lucky, divvy should show you the
> original division table. If you see the original division table. try
> 'mount -r /dev/u /mnt' to see if you can see the files. If you can
> see the u files, re-install UNIX on the other disk and recover the
> files by mounting the recovered disk's "u" division.
>
> 2) If divvy does not show the division table, you may have entered the
> wrong partition start and size, or the divvy table may be destroyed.
> If you are brave, run "mkdev hd 0 0" and setup a new division table
> with the information from your records for the divisions.
> >Be sure to "prevent" creation of new file systems on the divisions<
> as creating new file systems will destroy any hope of recovering your
> data. You must be able to run fsck on the "u" file system without
> getting an overwhelming number of errors. I would be happy with less
> then 10-20 un-referenced files. I would be suspicious I see a lot
> of DUP blocks and or unrealistic file sizes.
>
> 3) If you can successfully mount /dev/u from the emergency boot floppy
> set, reinstall UNIX on a new hard disk, then recover your files from
> the restored disk.
>
> NOTE: The above steps can be destructive if you make a mistake. You
> should purchase a new disk that is identical to the original 2 disks
> and use a disaster recovery product that will sector copy all data
> from one of the original disks to the "working" disk. Only work on the
> third disk. That way, if you butch the recovery on the new disk, you
> need only repeat the copy and try again.
>
> http://www.just-data-recovery-links.com/partition_partition_recovery_table_tool.html
>
> Good luck.
>
> P.S. Thanks for your post. I will print it and pass out copies to anyone
> trying to save money by omitting a good tape backup system on their
> UNIX server.