Messed up trying to migrate drive to RAID1

[IvanZD]

Hi,

I try to migrate from single drive to RAID1 setup, but it seems I screwed up data completely. Yeah, it looks like that. A hope still exist. Let see what we have:

I used this guide:
https://wiki.archlinux.org/index.php/Convert_a_single_drive_system_to_RAID

I copied data to degraded raid array, installed grub, rebooted few times into new degraded RAID system, everything appearently worked OK. Then I decided finally to add original drive to array. I did:

# sfdisk -d /dev/sdb | sfdisk /dev/sda

Checked with
# fdisk -l

to my eyes, everything was OK. I had now two identical drives, running on degraded raid 1 array.

Tried to add new drive to array, but I got:
# mdadm /dev/md0 -a /dev/sda2
mdadm: /dev/sda2 not large enough to join array

OK, guide said it is time to reboot if that happens. I issued reboot and after loading BIOS no go... it seems bad things happen.

[IvanZD]

Anoher info,

if I try in rescuecd:
# mdadm --assemble --scan -v

I got this for sdb drive (where half of md0 should be):
mdadm: cannot open device /dev/sdb: Device or resource busy


How is that?


Another strange thing:

In dmesg:
md: Autodetecting RAID arrays
md: invalid raid superblock magic on sda2
md: sda2 does not have a valid v0.90 superblock, not importing!
md: could not open unknown-block(8,18)
md: Scanned 2 and added 0 devices.

...

why now sda2 when working half of array should be on sdb ?

And also there is NO /dev/sdb1 and /dev/sdb2 at all! Only /dev/sda1 and /dev/sda2

whereas,

fdisk -l /dev/sdb lists both /dev/sdb1 and /dev/sdb2, the same as fdisk -l /dev/sda lists /dev/sda1 and /dev/sda2

[IvanZD]

Good news!

I recovered from this $%&/ by using testdisk utility on /dev/sda. It found all old partitions and wrote new partition table so I'm now again up and running on original disk. I was thought I won't see it again, but all data is fine thanks to this nice utility.

Now I'm gonna first backup all data to external disk and then play with RAID freely.

Thanks for reading!
_________________
http://www.meteoadriatic.net/

[IvanZD]

I want to open this thread again. Sorry for confusion but I need to solve this because it is important to continue.

Now when I booted again from original disk and tried to access /dev/md0 to see if it will work, it doesn't (expected), but somehow device /dev/md127 showed up and there is my (still degraded, on /dev/sdb2) RAID array, WORKING! What the... how, why? I can't explain this. Why md0 become suddenly md127?! Will this change randomly again and again?

And... now when I booted my original disk, I DO HAVE /dev/sdb1 and /dev/sdb2, that I didn't have in systemrescuecd as you can see in above posts.

Can anybody explain all this mess to me? Can this be related to new udev things? I have installed udev-171-r5. I don't know right now which one is on rescuecd I used. Maybe this explains the differences in behaviour between my system and rescuecd, but this still doesn't explain to me how suddenly md0 became md127?

Thanks.
_________________
http://www.meteoadriatic.net/