Fixing the /dev/md127 problem

A quick how to today.

/dev/md127. It happened again while I was adding a new RAID array to replace another one. You have no idea the number of times I had to look on the Internet on how to fix this problem. It usually take 1/2 hour to get the correct instructions, and I always tell to myself:

The solution to this problem is very simple. I know I will remember how to deal with it the next time.

But this is not the case, hence the reason for writing this post.

The fix is quite simple:

• Get the list of all raid arrays

  # mdadm --detail --scan
  

• Copy the array of interest from previous output. Remove the name variable and append what remains in /etc/mdadm.conf. If you have for example

  ARRAY /dev/md127 metadata=1.2 name=<hostname>:127 UUID=<some_uuid>
  

  just keep

  ARRAY /dev/md10 metadata=1.2 UUID=<some_uuid>
  

  Notice how /dev/md127 changed to /dev/md10 which is our desired device kernel name.

• Update the initramfs. In my case, on Parabola GNU/Linux-libre

  # mkinitcpio -p linux-libre
  # mkinitcpio -p linux-libre-lts
  

• Reboot and check if the RAID device name is now correct.

All this applies if you use mdadm and a GNU/Linux system.