Re: Notification for failed drive on hardware RAID controller

On Fri, 16 Feb 2007, Derrick Lobo wrote:
> Modern RAID controllers are convenient because they have their own BIOS
> that can build and maintain the attached drives into a RAID drive. The
> OS just has to see the drive and it's usable.
>
> The controller manufacturers have created software to run within the OS
> to manage these RAID drives. Versions for Windows have been around for
> quite awhile and I am now seeing versions that run in X as well.
>
> running a server that typicHowever, when ally has no windowed
> environment available, there is no software to manage the RAID and
> controller. Specifically no way to be alerted if a drive in the RAID
> pool fails.
>
> RAID is great, and an available hotspare can be configured to take over
> in case of a drive failure - but one still needs to be alerted so the
> failed drive can be replaced.
>
> Does anyone have an idea how to accomplish this?

The drives for hardware RAID controllers can report the current status (or 
at least, status changes) in dmesg, so you could grep this for suitable 
status information (for say, a nightly status email).

For proper management (e.g. to tell it to rebuild onto a spare drive), you 
will need management software for your RAID controller. NetBSD 3.1_STABLE 
and later include support for the FreeBSD tw_cli management tool for use 
with 3ware 8500 and 9500 series controllers (see the port-xen archives for 
a recent discussion of this). It's cli-based not X, but that's not a 
problem. I found it very easy to write a web-based interface onto it for 
use on one of our products, for example.

-- 
Stephen