mgb
My answers are assuming the drive is faulty.
Download the Ultimate Boot CD (UBCD) iso and create a bootable CD. Then use that to boot your sme server & run the manufacturers diagnostic test software to fully test and check the hard disk drive.
If the drive is faulty, then replace it & rebuild your system by installing a fresh sme OS and then restore from backup.
The exact approach will depend on how many drives you have in the sme server and whether they are in a RAID array.
Alternatively you can boot up the sme server using the sme install CD and select the rescue mode option, or IIRC type sme rescue
You can then use Linux tools to interrogate/test the problematic drive.
With a failing drive, the theory is to stop using it as further disk accesses may further corrupt the data on the disk. Ideally you would connect and mount the faulty disk in another sme server, and when mounted, you may be able to copy valuable (recent) data off the drive to a known good drive, for later manual restore to a rebuilt sme server.
The other theoretical approach is to mount the drive in another sme server and then use the dd command to copy every bit off the apparently faulty drive to a known good clean drive, ie three drives involved here, a good drive with sme OS, the second mounted faulty drive to copy from and a third good mounted drive to copy to.
At the command prompt of another sme server (or google it) do
man dd
for examples or search these forums for a suitable command syntax.
Remember it will take many hours to do a dd copy of every bit on your hard disk, depending on the size. This way you will have a copy of everything that is still readable on your failing drive, but on a physically good drive, so the chances of data recovery are greater.
If you have known good backups that are recent enough to use, then you can skip the disk copy process.