beginner error detection

beginner error detection

[Tomka Gergely]

Hi!

I have a simple sw raid1, over two sata disks. One of the disks started to 
complain (s.m.a.r.t. errors). I think in the near future i witness a disk 
failure. But i don't know how this thing is happening with raid1, so i 
have some questions. If these questions answered somewhere (faq, manpage, 
url), then feel free to redirect me to this source(s). 

Can the raid1 detect and handle disk errors? If one block goes wrong, how 
can the raid1 driver choose which was the correct, original value?

Sata systems can die gracefully? When in a good scsi system happens a 
total disk failure, then the scsi makes the disk fail, mdadm removes 
the disk from the array, and in the morning i see a nice e-mail. When a 
PATA disk dies, the system goes down, so i need to call a cab. I dont know 
how the sata behaves in this situation.

The kernel 2.6.20 (with skas patch), the controller :
nVidia Corporation CK804 Serial ATA Controller (rev f3)

Thanks.

-- 
Tomka Gergely, gergely@tomka.hu

Re: beginner error detection

[Mario 'BitKoenig' Holbe]

Tomka Gergely <gergely@tomka.hu> wrote:
> Can the raid1 detect and handle disk errors?

This is what it is designed for :)

> If one block goes wrong, how 
> can the raid1 driver choose which was the correct, original value?

RAID1 depends on the disk to state errors. RAID1 does not ensure data
integrity. RAID1 cannot correctly handle a disk returning wrong data
(i.e. other data than written before). Only RAID2 is designed for such.
However, since todays disks have their own data integrity detection
(i.e. ECC), they should normally not return wrong data but read-errors
instead.

And assuming a drive returns either correct data or a read-error, it is
simple to choose which was the correct, original value :)


regards
   Mario
-- 
> As Luke Leighton said once on samba-ntdom, "now, what was that about
> rebooting?   that was so long ago, i had to look it up with man -k."