Vital Patches for ataraid with Intel Matrix RAID (ICH7)

[Jeremy Chadwick]

On Wed, Mar 26, 2008 at 11:00:49PM -0700, Jeremy Chadwick wrote:
> On Sat, Mar 22, 2008 at 04:58:47AM +0000, Stef Walter wrote:
> > Daniel O'Connor wrote:
> > > I have seen this bug in other ATA RAID implementations (VIA & Promise) 
> > > too. From what I can tell this part of your patch is general to all ATA 
> > > RAID arrays, right?
> > 
> > Yes, a small part. The part that will write out the RAID information
> > (thus updating the generation) whenever the status changes, regardless
> > of whether that change takes place when the machine is off or not.
> > 
> > Without testing, I can't be sure whether this solves the problem on
> > other ataraid devices.
> 
> I have access to many boxes that have two disks, support native Intel
> MatrixRAID, and proper hot-swap backplanes.  I also have access to a
> Promise TX4310 card, which I can hook up to said servers (vs. using the
> onboard ICH7).
> 
> I should be able to reproduce failure scenarios (pulling disks while in
> a RAID-1 configuration, etc.) on those systems and report back here.
> 
> Would that be ideal?

Off-topic: Stef, if you read mailing lists: the MX associated with
memberwebs.com is still using ORDB.  You need to contact said mail
administrators can tell them to stop using it, because it's returning
false positives for every mail delivered to that domain:

<stef at memberwebs.com>: host memberwebs.com[209.66.100.222] said: 554 Service
    unavailable; Client host [72.20.106.3] blocked using relays.ordb.org;
    ordb.org was shut down on December 18, 2006. Please remove from your
    mailserver. (in reply to RCPT TO command)

-- 
| Jeremy Chadwick                                    jdc at parodius.com |
| Parodius Networking                           http://www.parodius.com/ |
| UNIX Systems Administrator                      Mountain View, CA, USA |
| Making life hard for others since 1977.                  PGP: 4BD6C0CB |