Subject: Re: RAIDframe: notify me when a drive fails?
To: Geert Hendrickx <firstname.lastname@example.org>
From: Greg Oster <email@example.com>
Date: 02/28/2006 13:46:52
Geert Hendrickx writes:
> Content-Type: text/plain; charset=us-ascii
> Content-Disposition: inline
> On Tue, Feb 28, 2006 at 06:21:02PM +0100, Geert Hendrickx wrote:
> > It doesn't have to be a separate daemon per se. Something I can (easily)
> > check for via a cron-job is ok. The grep-for-failed thing works, but it
> > would be more elegant if e.g. raidctl -s would return an exit status >0 if
> > something is wrong and needs human intervention. Maybe combined with a -q
> > flag (for no output), it would get as easy as
> > raidctl -s -q raid0 || mail -s "RAID problem" ...
> > in an hourly cronjob.
So how about just:
raidctl -s raid0 | grep failed && mail -s "RAID problem" ...
? (Works fine in the testing I just did :) )
> This simple patch (against 3.0) makes "raidctl -s" return the number of
> failed components and/or spares in the array (i.e., normally 0).
I really don't care for mixing of error codes and "number of disks that
have failed" like that... :( (never mind that it wouldn't let
raidctl return real error codes for other legitimate reasons at some