Re: Server down-fail​ed RAID5-asking for some assistance

[NeilBrown <neilb@suse.de>]

On Sun, 24 Apr 2011 12:57:56 +0100 John Robinson
<john.robinson@anonymous.org.uk> wrote:

> On 24/04/2011 09:41, NeilBrown wrote:
> > On Sun, 24 Apr 2011 01:06:52 -0600 John Valarti<mdadmuser@gmail.com>  wrote:
> [...]
> >> Here you go. I warn you it is a bit long!
> >
> > Thanks.  Length is no problem.
> >
> > Only it doesn't make sense at all.  I cannot see how mdadm would possibly be
> > generating just those messages.
> > It appear to be rejecting each device for some reason, but it not reporting
> > why it is rejecting the device...
> >
> > What version of mdadm is this?
> >     mdadm --version
> >
> > I should have asked that before.
> 
> I think John said 2.6.9


Ahh, I see it.  This is a bug in there:  ->used isn't set to zero after 'dv'
is allocated.  This was fixed in 3.0.  I don't remember that bug...

I cannot see any easy way to work around that bug.
You could possibly:

  echo DEV /dev/sd[abc]2 > /tmp/mdadm.conf
  mdadm -Eb /dev/sda2 >> /tmp/mdadm.conf
  mdadm -Afvv /dev/md1 -c /tmp/mdadm.conf

I think that would work - but no promises.

 on the CentOS 5.5 rescue media. I think it's 
> time to try something more recent: John, could you try SystemRescueCD 
> from http://www.sysresccd.org/ and run
>    mdadm -Evvs
> and if that shows your RAID5 members again,
>    mdadm -Afvv /dev/md1

Getting a newer mdadm is definitely a good idea.

Safest to explicitly list the devices that you want
     mdadm -Afvv /dev/md1 /dev/sd[abc]2


NeilBrown

> 
> Cheers,
> 
> John.