[NotarySojac]

So I've successfully setup a software RAID 1 using mdadm across two hard drive partitions. Then just for kicks I decided to unplug one of the drives and see if the data was still fine. It was! Great, but then when I plugged in the drive again, it was first detected as /dev/sdc instead of what it used to be /dev/sdb. I didn't want to confuse my system so I decided to reboot, and it was picked up as /dev/sdb proper, but mdadm didn't seem willing to enter it back into the RAID.

I think it was telling me...

$  cat /proc/mdstat
Personalities : [raid1]
md0 : active raid1 sda1[0]
      1464845671 blocks super 1.2 [2/1] [U_]  

I tried lots of stuff, but couldn't get it to be an 'active' [UU] raid again... stuff like...

$  mdadm --assemble /dev/md0 /dev/sda1 /dev/sdb1
mdadm: /dev/md0 assembled from 1 drive - need all 2 to start it (use --run to insist).

The bottom of my /etc/mdadm/mdadm.conf looked like this:

ARRAY /dev/md0 metadata=1.2 name=Mycomp:0 UUID=6e7efcee:e9bc12a6:cb210f4f:0f0b4736

Which I believe was generated by $ mdadm --detail --scan >> /etc/mdadm/mdadm.conf.


Now I'm re-mirroring the drives... (ughh, such a pain...) and starting over, to make sure it was originally a [UU] raid, and it seems like it is in fine condition, but now, how do I perform this 'drive unplugging' test and rebuild the array properly? Am I missing a step/ conf configuration?

[NotarySojac]

I finally decided to just reattempt the test and the second time, it worked without a glitch. I'm not sure what could have caused it in the first place. I thought I was going to need to do something more technical like http://www.ducea.com...dm-cheat-sheet/ but that wasn't the case.

[no2pencil]

Glad to see that you got this working!

In the past when I spanned a volume across multiple SCSI drives, if there was an unsuspected power outage that crashed the system, I would have to clean the file system on the drives before I could address them as a single block device. & with no block device, there is nothing to mount