Degraded raid handling

[Heinz Mauelshagen]

On Mon, May 21, 2007 at 05:57:49PM +0200, Tormod Volden wrote:
> On 5/21/07, Phillip Susi <psusi at cfl.rr.com> wrote:
> >It could happen that the bios sees both disks just fine, but for
> >whatever reason, the kernel only sees one.  In that case, dmraid will
> >activate the set with just the one disk, but without updating the
> >metadata to indicate this has been done, the set will be corrupted at
> >the next boot.  Not good...
> 
> Already happened here, I think... I have my two disks mirrored on an
> "Intel Corporation 82801GR/GH (ICH7 Family) Serial ATA Storage
> Controller AHCI". From the grub prompt I can browse (hd0,2) and
> (hd1,2) and I can see there's a difference between them. The BIOS tool
> says everything is fine (and there's no option to repair).

Hrm, that sounds like a BIOS flaw to me.

> 
> The damage was probably caused either by playing with udev rules or by
> having savedefault in the grub menu.lst.
> 
> $ sudo dmraid -s
> *** Group superset isw_dhedigiiib
> --> Active Subset
> name   : isw_dhedigiiib_osmoraid
> size   : 488390656
> stride : 256
> type   : mirror
> status : ok
> subsets: 0
> devs   : 2
> spares : 0
> 
> $ sudo dmraid -r
> /dev/sda: isw, "isw_dhedigiiib", GROUP, ok, 488397166 sectors, data@ 0
> /dev/sdb: isw, "isw_dhedigiiib", GROUP, ok, 488397166 sectors, data@ 0
> 
> $ sudo dmraid -b
> /dev/sda:    488397168 total, "S09QJ10YB19706"
> /dev/sdb:    488397168 total, "S09QJ10YB19699"
> 
> $ sudo dmraid -tay
> isw_dhedigiiib_osmoraid: 0 488390656 mirror core 2 131072 nosync 2
> /dev/sda 0 /dev/sdb 0
> ERROR: opening "/dev/.static/dev/mapper/isw_dhedigiiib_osmoraid"
> 
> What does "nosync" mean? I get the same on another machine where I
> thought everything was fine though.

It means that the mirrors in the set are not going to be synced by the
device-mapper mirroring code.

> 
> However, I wonder if I screwed up when I chose AHCI in the BIOS before
> installing - I thought it meant ahci+raid - the OS'es think I am
> running RAID, but I am not sure about the BIOS. (XP sees only the raid
> volume so it seems to work fine, but the Intel Management Console does
> not say anything about RAID, just lists the two disks. And grub sees
> two disks.)

Well, there's obvious metadata left behind telling that this is
a RAID set. If the BIOS doesn't detect it, its RAID capability is probably
turned off.
dmraid relies on metadata on any disks only and discovers such.
It doesn't pay attention to any HBA specifics, just metadata.

Heinz

> 
> Tormod
> 
> _______________________________________________
> Ataraid-list mailing list
> Ataraid-list at redhat.com
> https://www.redhat.com/mailman/listinfo/ataraid-list

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-

Heinz Mauelshagen                                 Red Hat GmbH
Consulting Development Engineer                   Am Sonnenhang 11
Storage Development                               56242 Marienrachdorf
                                                  Germany
Mauelshagen at RedHat.com                            PHONE +49  171 7803392
                                                  FAX   +49 2626 924446
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-