Accidentally nuked my system - any suggestions ?


My workstation is running CentOS 7 on two disks (sda and sdb) in a
software RAID 1 setup.

It looks like I accidentally nuked it. I wanted to write an installation
ISO file to a USB disk, and instead of typing dd if=install.iso
of=/dev/sdc I typed /dev/sdb. As soon as I hit <Enter>, the screen froze.

I tried a hard reset, but of course, the boot process would stop short
very early in the process.

Now, I have backups of the important stuff of course, so no real
catastrophe. But it would be nice if I could get back the data from my
disk directly.

I booted a rescue disk (Slax 9.6.4) and I can see my disks as well as
raid arrays /dev/md125, /dev/md126 and /dev/md127. Oh, my partitioning
scheme is manual and quite simple. Everything is RAID 1, I have a /boot
array on /dev/sda1 + /dev/sdb1, swap on /dev/sda2 + /dev/sdb2 and / on
/dev/sda3 + /dev/sdb3.

I tried to mount /dev/sda3 directly from the rescue disk:

# mount /dev/sda3 /mnt

But I only get this:

mount: unknown filesystem type 'linux_raid_member'

I'd be very grateful for suggestions.




Re: Accidentally nuked my system - any suggestions ?

By Gordon Messmer at 12/04/2018 - 17:10

On 12/4/18 2:01 PM, Nicolas Kovacs wrote:

The system should boot normally if you disconnect sdb.  Have you tried that?

Re: Accidentally nuked my system - any suggestions ?

By Niki Kovacs at 12/04/2018 - 17:31

Le 04/12/2018 à 23:10, Gordon Messmer a écrit :
Unfortunately that didn't work. The boot process stops here:

[OK] Reached target Basic System.

Now what ?

Re: Accidentally nuked my system - any suggestions ?

By m.roth at 12/04/2018 - 17:13

Gordon Messmer wrote:
Duh! thanks, Gordon, a simpler answer than mine, with the same effect,
that /dev/sdb failed as far as mdadm was concerned.