How about this method.

Usually alle servers can spare a backup partition of about 300 MB.
Install a clean Os and the *SM client on this partition.
Give the backuppartition the same netbios and nodename as the original
partition.
Edit boot.ini so you can access the backup partition after you reboot the
corrupted system.

Let's say the backup partition is drive E: and the system and data disks are
C: and D:.
Off course your system has been succesfully backupped previously while
having the systempartition on disk c: active.

For some reason the OS gets corrupted.
Reboot the server to the backuppartition.
From here you can restore the complete systemdisk C: and datadisk D:
Decide on the existing data, whether or not to use replace=yes.

The good thing about this backup partition, is that the systemdisk never has
files in use, while doing a bare metall restore.
Restore and reboot back to the original partition, and all is well.
Over here this works fine.
The only requirement is that you have a spare partition of about 300 MB to
use for backup partition.



Met vriendelijke groet,
Kind regards,

___________________________________________________

Ilja G. Coolen
ABP / USZO / CIS / ES / Storage Management
Tel: +31 (0)45 579 7938
Fax: +31 (0)20 882 7286
E-mail : ilja.coolen AT abp DOT nl <mailto:ilja.coolen AT abp DOT nl>