Reboot into emergency mode caused by RAID and getting rid of /dev/md0


We have a 3 x 2TB disk and didn’t want to use software raid, so, we got rid of /dev/md0 by mdadm --remove /dev/md0. We already have successfully mounted all individual ssd and works ok…

In any rate, after installing CUDA 9, and reboot, the system goes into emergency mode and not boot up any more. systemctl --failed does not show anything have failed.

What else can I try to make it boot? Is there more to remove raid setup completely?

System detail:
/dev/sda - Ubuntu is installed, fsck check states no problem exists.

These 3 ssd disk is no longer a mirrored disk (was mounted as /data). After mdadm --remove, I was able to mount it separately - which is what we want.


Hi Jae,

My guess as to why your system is going into emergency mode would have to do with an /etc/fstab entry for /dev/md0 that still remains even after you removed the mdadm raid.

It’s likely that the output of mount -a is:

$ sudo mount -a
mount: special device /dev/md0 does not exist


$ sudo mount -a
mount: can't find UUID=27f957a4-a623-4e1a-956f-9df2bcca8e75

(The UUID will be different and will be the UUID for the old Linux MD RAID.

Please do the following to fix this:

$ sudo nano /etc/fstab
# Delete or comment out the line that looks like:
/dev/md0 /data ext4 defaults 0 0
# or
UUID=17f957a4-a623-4e1a-956f-9df2bcca8e75 /data ext4 defaults 0 0
# You can add a comment by adding a # to the beginning of the line
# Write the changes to disk
# CTRL + o then ENTER then CTRL + x to 

Then run sudo mount -a again and it should exit with no output. You should then be able to freely reboot with no emergency mode errors.