Re: accidently pulled to many devices, raid6 wont start.

[Bernd Schubert <bernd.schubert@fastmail.fm>]

On 12/05/2013 03:20 PM, Wilson Jonathan wrote:
> On Thu, 2013-12-05 at 14:36 +0100, Bernd Schubert wrote:
>> On 12/05/2013 01:30 AM, Wilson Jonathan wrote:
>>> mdadm: /dev/sdf6 is identified as a member of /dev/md5, slot 2.
>>> mdadm: /dev/sde6 is identified as a member of /dev/md5, slot 1.
>>> mdadm: /dev/sdd6 is identified as a member of /dev/md5, slot 0.
>>> mdadm: /dev/sdb6 is identified as a member of /dev/md5, slot 4.
>>> mdadm: /dev/sda6 is identified as a member of /dev/md5, slot 3.
>>> mdadm: ignoring /dev/sde6 as it reports /dev/sdf6 as failed
>>
>> [...]
>>
>>> So tried again, with a different (valid) chunk...
>>>
>>>
>>> root@BorgCUBE:/mnt/datastore/wilsonjonathan# mdadm --create
>>> --assume-clean --level=6 --raid-devices=6
>>> --chunk=64 /dev/md5 /dev/sdd6 /dev/sde6
>>> missing /dev/sdf6 /dev/sda6 /dev/sdb6
>>
>> Why are you using this order? Missing seems to be at the wrong place?
>>
>>
>> Cheers,
>> Bernd
>
> Arr yes, I seem to have got the order wrong... before I continue and do
> further damage; I put the original pulled drive (sdc) back in and re-ran
> examine. Obviously I have corrupted the raid layout data on the other
> disks (a,b,d,e,f) but here is the sdc examine output.
>
> /dev/sdc6:
>            Magic : a92b4efc
>          Version : 1.2
>      Feature Map : 0x1
>       Array UUID : 9ff95b8b:ba34ad95:aa7cb806:169246f2
>             Name : PartedMagic:6
>    Creation Time : Fri Nov 25 16:30:19 2011
>       Raid Level : raid6
>     Raid Devices : 6
>
>   Avail Dev Size : 1833569679 (874.31 GiB 938.79 GB)
>       Array Size : 3667138560 (3497.26 GiB 3755.15 GB)
>    Used Dev Size : 1833569280 (874.31 GiB 938.79 GB)
>      Data Offset : 2048 sectors
>     Super Offset : 8 sectors
>            State : active
>      Device UUID : 4ce632fb:d506da60:120cac8e:4433efc4
>
> Internal Bitmap : 8 sectors from superblock
>      Update Time : Wed Dec  4 23:04:40 2013
>         Checksum : fe4e7b1f - correct
>           Events : 260243
>
>           Layout : left-symmetric
>       Chunk Size : 64K
>
>     Device Role : Active device 5
>     Array State : AAAAAA ('A' == active, '.' == missing)
>
>
>>
>>
>
> I mistook the order (a-b-c-d-) with the position number in the array.
>
> If I'm reading this correctly, Active device 5 means it should be at
> position 5, or the sixth disk in the array.

Yes, and the intial kernel output also suggest that.

>
> I also note it has a bitmap on it, something the other drives now lack
> as it was not stated in the original forced create, so not sure how that
> will affect things and the event count is lower, 260243, than the other
> devices, 260321, before the array became corrupted (that said I know
> exactly which files were being updated on the file system, so if I can
> get it back I will just delete them and re-create).
>
> Hopefully I have not totally killed the array by my mistakes.

I would simply try again with the correct order - 'missing' as last 
argument. Then check your data, i.e. test if it mounts (read-only) and 
run fsck (read-only).

I'm also a bit suprised why you went the '--create' way at all, why 
didn't you simply try to assemble with 4 drives only?


Cheers,
Bernd