can't activate raid0 array after kernel upgrade

[raginaot]

Hi
I've been trying everything I can to solve this problem, but so far no luck.
I am a total newb at this though, only been using linux as a main OS for a few weeks.

I'm running Arch linux, kernel 2.6.36 (it was working fine on 2.6.33)

My raid array is on the onboard intel controller (Gigabyte EX58-UD3R mobo). It also has a GSATA controller which I have tried as well but it's the same problem with that one.

It's a 2 disk raid0 array and I use a ~300GB partition of it as a system drive for my Windows 7 installation.
I can boot from the array and windows works flawlessly on it, but as soon as I boot into linux it gets all messed up. I can't activate with dmraid or mdadm and when I reboot from linux the array FAILS in the raid bios showing only one drive as a part of the array. (both drives are detected though, but only one of them shows "raid-active").

It starts to work again after cold-boot though, which is pretty weird to me.
It's always the same disk that fails too.

I included some info from my kernel log and fdisk

the disks appear like this

Disk /dev/sdd: 1000.2 GB, 1000204886016 bytes
255 heads, 63 sectors/track, 121601 cylinders, total 1953525168 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0xb876c595

Disk /dev/sdd doesn't contain a valid partition table

Disk /dev/sde: 1000.2 GB, 1000204886016 bytes
255 heads, 63 sectors/track, 121601 cylinders, total 1953525168 sectors
Units = sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x1c7fba32

   Device Boot      Start         End      Blocks   Id  System
/dev/sde1   *        2048      206847      102400    7  HPFS/NTFS
/dev/sde2          206848   552962047   276377600    7  HPFS/NTFS
/dev/sde3       552962048  3907033087  1677035520    7  HPFS/NTFS

and here's a few lines from my kernel.log that I thought might be relevant.

Dec  6 05:04:54 behemoth kernel: sde: sde1 sde2 sde3
Dec  6 05:04:54 behemoth kernel: sde: p3 size 3354071040 extends beyond EOD, enabling native capacity
Dec  6 05:04:54 behemoth kernel: ata9: hard resetting link

and

Dec  6 05:04:54 behemoth kernel: sd 8:0:0:0: [sde] 1953525168 512-byte logical blocks: (1.00 TB/931 GiB)
Dec  6 05:04:54 behemoth kernel: sde: detected capacity change from 1000203804160 to 1000204886016
Dec  6 05:04:54 behemoth kernel: sde: sde1 sde2 sde3
Dec  6 05:04:54 behemoth kernel: sde: p3 size 3354071040 extends beyond EOD, truncated
Dec  6 05:04:54 behemoth kernel: sd 8:0:0:0: [sde] Attached SCSI disk

you can see the entire log here - pastebin.com/duWu8zYB

now I added this to my mdadm.conf, not sure if it's right though.

ARRAY /dev/md0 devices=/dev/sdd2,/dev/sde2

but when I do "mdadm --assemble --scan"
I get the following output

mdadm: /dev/sde2 has no superblock - assembly aborted

All I can think of is that it seems like the system is reading the partition table of /dev/sde as it would a single non-raided disk and therefore it looks like the partition is bigger then the disk it self so it looks like its failing.

But I don't know how I can fix it.

If someone could point me in the right direction or even explain some of those things in the kernel log it would great.

Thanks in advance

--raginaot

[Manko10]

Have you added that disk before with

mdadm --add /dev/md0 /dev/sde2

If /dev/sde2 has never been part of the array before it doesn't have a RAID superblock.

[raginaot]

Have you added that disk before with

mdadm --add /dev/md0 /dev/sde2

If /dev/sde2 has never been part of the array before it doesn't have a RAID superblock.

I have not, and I can't since appearantly md0 doesn't exist
But I never tried mdadm before I started having problems, I always used dmraid.
Just did "dmraid -ay" and it was automatically activated and the partitions were created as /dev/dm-2 and /dev/dm-3 so I could mount them.
If I try that now I get:

/dev/sdd: "jmicron" and "isw" formats discovered (using isw)!
ERROR: isw: wrong number of devices in RAID set "isw_ceijhcahbb_TERARAOD" [1/2] on /dev/sdd
RAID set "isw_ceijhcahbb_TERARAOD" was not activated
ERROR: device "isw_ceijhcahbb_TERARAOD" could not be found

But I tried creating a new array using

sudo mdadm --create /dev/md0 -l0 -n2 -c128 /dev/sde2 /dev/sdd2

but the output I get is

mdadm: super1.x cannot open /dev/sdd2: No such file or directory
mdadm: /dev/sdd2 is not suitable for this array.
mdadm: create aborted

so now it's having problem adding sdd2 :S
am I doing it wrong?

But like I said in the first post, as soon as I boot into linux, the second disk in the array is deactivated in the raid BIOS so I'm certain there is some underlying problem here.

I don't know, I'm gonna get some sleep and hack around some more in the "morning" :P


edit: stupid me, I seem to have permanently borked the array somehow. Thank god for backups...
I created a new one and now mdadm seems to be automatically creating this device on boot called /dev/md127 1717.3GB of size and with no partition table.

[Manko10]

Indeed, you use the fake hardware RAID of your motherboard. That's always delicate.

To rebuild your array you should go with dmraid -R. Something like that:

dmraid -R isw_ceijhcahbb_TERARAOD /dev/sde2

Maybe you also have to play around with your BIOS settings a bit. You can find more information here: [opensuse] Removing a failed disk from a dmraid array
But I'm not an expert here at all so I won't be able to give you further advice if that doesn't work.

[raginaot]

IT WORKS

I'm not entirely sure what I did, but I started by permanently ****ing up the array with some mdadm command, then I deleted it in bios, then I had to delete all the partitions in another windows machine cause the table was so majorly fubar linux thought one of the drives was several PB in size :S
Then I created a new array, a few reboots, banged my head on the keyboard for a while and sent Linus Torvalds an angry letter and somehow it's working now.

Cheers for trying to help manko. appreciate it