[Beowulf] MD check/scrub

Leif Nixon nixon at nsc.liu.se
Tue Nov 13 09:03:11 PST 2007


Mark Hahn <hahn at mcmaster.ca> writes:

>>> not me.  raid is too important to be trusted to hardware - have you
>>> tried MD's check/scrub features?
>>
>> Yes, now. Or rather, Peter Kjellström has.
>>
>> If a raid 6 stripe has a corrupted block, it seems md "repairs" the
>> damage by recalculating the parity blocks. Less than ideal...
>
> pardon my density, but what is the alternative?

Reconstruction. With raid 6, you can recover from single-disk
corruption (As opposed to *failures*, where you get read errors from a
disk. Raid 6 can handle two simultaneous disk *failures*.).

See section 4 in:

http://www.kernel.org/pub/linux/kernel/people/hpa/raid6.pdf

Just recalculating the parity blocks does give you a consistent raid
stripe, but destroys your data (unless it actually was one of the
parity blocks that was corrupted).

-- 
Leif Nixon                       -            Systems expert
------------------------------------------------------------
National Supercomputer Centre    -      Linkoping University
------------------------------------------------------------




More information about the Beowulf mailing list