[clug] Is this logic correct? [12Tb RAID-5 array unlikely to rebuild]

[Andrew Janke]

On Sat, Mar 28, 2009 at 19:32, Brett Worth <brett at worth.id.au> wrote:
> steve jenkin wrote:
>> Anyone care to comment if this logic is correct or not?
>> (Large drives in RAID-5 likely to be unable to rebuild)
>> Seems simplistic to me...
>>
> In recent years I've had it happen to me twice on RAID5 volumes that I got a read error
> during reconstruct and lost data i.e. had to go to backups.  The drives were 400 and 250GB
> SATA.

Had it happen once.  (16x 750GB SATAII in a Chenbro RM414 case with an
Areca Controller).  Although it did manage to still rebuild after a
second hack through.  Guess I was lucky. :)  Mind you I was only using
RAID5 as the new RAID6 cards hadn't landed yet.   Now I have (at last
count) 15 or so of these systems scattered about the globe running
RAID6 with two hot spares on both Areca and 3ware 9500 S series cards
with no trouble so far.

Both the cards above support SMART monitoring but like others there is
not always a correlation between what SMART tells you and what comes
to bear.... In any case the systems are now all redundant (there is a
second machine that mirrors the first continuously) so I sleep OK.

YMWV but I am happy with cheap disks in SATAII arrays, throughput can
choke at times if there are a bazillion processes (usually nfs)
writing at the same time but by en large very good bang for buck.


--
Andrew Janke
(a.janke at gmail.com || http://a.janke.googlepages.com/)
Canberra->Australia    +61 (402) 700 883