I think I messed up pretty bad. On my old system, I had 2 identical drive that were assembled with mdadm as a MD RAID 10 with a "far" layout.

When building my new system and migrating the two disks, I had a total brain fart, and instead of doing assemble and scan, I ran the following command:

When /dev/md0 wouldn't mount, I quickly realized that this was not what I intended, but the damage was done, as mdadm had started the sync. I have stopped mdadm (mdadm --stop /dev/md0), and now I can't mount the drives individually (unknown filesystem type 'linux_raid_member'), and parted doesn't see any filesystem on the partitions of the individual drives.

Is there a way to recover the data from either drive, or did mdadm completely destroy both drives in an unrecoverable way?

I'm sad to say I think its probably hosed the data due to the re-sync... although that said, it "might" be possible to recover some of the data further into the partition (which escaped the sync) it just depends on if the important file system inode data is still hanging around some where further along on the disk, if the file "name/inode" data (as apposed to the data of/in the file) is available some data might be recoverable, but I'll be honest I think its unlikely.

Your best bet is to use the mailing list for linux raid. http://vger.kernel.org/vger-lists.html << the linux raid list is about 2/3's the way down the page.

If you've never used mail lists before, they are fairly simple but you'll need to do a little reading first.