philo wrote:
On 11/03/2015 07:26 PM, Paul wrote:


I bet it has a smokin transfer rate.
Probably good for 4-5MB/sec.


I did not bench mark the drive but it ran msdos 6.22 just fine.


I save small hard drives for older machines that I put Linux on.


Now another question:


The drive had a bad sector that was mapped out by scandisk.


If I would re-use the drive for dos or Windows, I assume the OS would
not try to use that sector...but does scan disk map it out just for
ms-based operating systems...or is it mapped out at a lower level?

My question is if I used the drive for Linux would that bad sector cause
a problem?

My guess is that scan disk only maps it out for MS operating systems.


What I did was use both test disk and d-revitalize to attempt to make
that bad sector invisible to any OS.


I again ran scandisk and it shows no bad sectors.

Can I assume it's been mapped-out at a lower lever than the OS?


At the sector level, the disk has automatic sparing, as a
feature of the IDE controller. If a bad sector is detected,
a spare is used in its place.

If the drive runs out of (local) spare sectors, the sector
is then un-repairable. The drive returns "CRC error" when the
sector is read.

Now, it's up to the file system to mark it. The file system
uses a sparse file, which keeps track of sectors or clusters
with CRC errors. That's how it would take a bad area out
of circulation. The "real space on disk" taken by the sparse
file, grows with the more bad areas it needs to map. Even though
the virtual size of the file is as large as the disk, space is only
actually used to store that file, when there is a need to mark an
area of the disk as bad.

You would need to research the various file system types,
to see what mechanisms they use.

https://en.wikipedia.org/wiki/Fat

Bad blocks Cluster tagging

https://en.wikipedia.org/wiki/Ntfs

Bad blocks $BadClus (MFT Record)

https://en.wikipedia.org/wiki/Extended_file_system

Bad blocks Table

So it looks lkke they all have some sort of mechanism.

Another way to handle the issue (for the simplest cases),
is to move the partition so it does not sit on top of the
bad area.

I had one disk, where all the "trouble" was in a single
well defined area. For that disk, it would be feasible to
move the partitions away from there. And then I would
only lose around 14% of the drive capacity.

Paul