extrapolate data from RAID

  • Thread starter Thread starter Kirk
  • Start date Start date
K

Kirk

Does anyone know if it's possible to extrapolate data from a single
drive that was part of a large RAID set? RAID 0? RAID 4? RAID 5?

Thanks..
 
Maybe if the file size is smaller than the stripe size, but I have never
done it myself. Really just a guess.

--Dan
 
Previously Kirk said:
Does anyone know if it's possible to extrapolate data from a single
drive that was part of a large RAID set? RAID 0? RAID 4? RAID 5?
Click to expand...

RAID0: You get exactly half the data sectors.

RAID4: Exotic. If you get a non-parity disk, you get 1/n of the
data sectors. With the parity disk you get the same number
of sectors with (n-1) data sectors XORed together.
Whether this has any value depends entirely on the
data on the disks.

RAID5: You get 1/n (n drives) of the data sectors reliably and
a further 1/(n*n) XORs of (n-1) data secors. Whether these
have any value depends on the data on the disks.

So no, there is no possibility for interpolation.

But you can get a fair amount of data direct and some partial
information about other parts of the data as well.

Arno
 
Wrong again.
Click to expand...

Hey Folkert, while I appreciate the attention, you seem to have
lost all will to distinguish my wrong statements (and they exist)
from the things were I know what I am talking about.

Maybe you should try to moderate this obsession of yours? This
cannot be healthy. Maybe go out with some folks and just
relax for a time?

Arno
 
Arno Wagner said:
RAID0: You get exactly half the data sectors.
Clueless.


RAID4: Exotic. If you get a non-parity disk, you get 1/n of the
data sectors. With the parity disk you get the same number
of sectors with (n-1) data sectors XORed together.
Whether this has any value depends entirely on the
data on the disks.
Nope.


RAID5: You get 1/n (n drives) of the data sectors reliably and
a further 1/(n*n) XORs of (n-1) data secors. Whether these
have any value depends on the data on the disks.
Click to expand...

Wrong again.
 
Back
Top