[ILUG] Raid disk query

[kevin]

There is a line in mdadm.conf that says 

DEVICE partitions /dev/sda1 /dev/sda2

So I guess its old device names.

I suspect the answer is what was mentioned earlier about using desktop
drives in raid. I have seen the problem twice now on two different
arrays and both problems happened immediately after reboot. When the
drives are reintroduced into the array they work away fine with no
problems.

It would be handy to be able to configure the raid array to be a bit
slower in dumping out drives.

Kevin.

-----Original Message-----
From: FRLinux <frlinux at frlinux.eu>
To: kevin <lists at cybercolloids.net>
Cc: ilug <ilug at linux.ie>
Subject: Re: [ILUG] Raid disk query
Date: Thu, 17 Nov 2011 10:52:40 -0000

On Fri, November 11, 2011 12:28 pm, kevin wrote:
> I rebooted one of our servers yesterday and after rebooting mdadm
> complained that one of the raid disks (raid 1 array, two operational and
> one spare) had failed and it began to resync to the spare disk.
>
> The question is - why? SMART doesn't report anything wrong with the
> disk.
>
> I am loathed to dump a good disk - any suggestions?

Is your raid declared by old device naming scheme such as /dev/sda or is
it using uuid? (/dev/disk/by-uuid)?

I could foresee a boot issue where the second drive is not caught up in
time and fails your RAID. Happened a lot on the crappy sil3112
controllers.

Cheers,
Steph