Re: [tlug] How to proceed with RAID drive failure

[Attila Kinali <attila@example.com>]

On Thu, 14 Jul 2011 09:50:35 +0200
Josh Glover <jmglov@example.com> wrote:

> > [I]t's possible that just plugging in
> > the drive in the same place would cause it to be considered part of
> > the array and reintegrated unless the BIOS was reset.  Be careful!
> 
> That is terrifying!

That's actually what happens if you have a "real server [tm]"
(like a HP DL or IBM x series). There, all disks in the server
are managed by the RAID system. There is no way to access a disk
directly. This means that if you partitioned and formated a disk
outside and wanted to take the data over, you'd lose it all. But
as you cannot access the data anyways, you would not try it at all ;-)

These systems have the big advantage that you get a red light which
tells you which disk to change (no takinging-out-the-wrong-disk-of-the-
raid-and-screwing-the-whole-system-that-even-the-backup-doesn't-help-anymore)
and just placing a new disk triggers a sync and gets system back to normal.
Thus you can use a janitor to walk around in big data centers, looking
for red lights and switching disks as needed. No other interaction needed.

			Attila Kinali

-- 
The trouble with you, Shev, is you don't say anything until you've saved
up a whole truckload of damned heavy brick arguments and then you dump
them all out and never look at the bleeding body mangled beneath the heap
		-- Tirin, The Dispossessed, U. Le Guin