Warning: mdadm 4.5 in Omarchy 3.4.0

[jonttsw]

mdadm now requires kernel module md_mod to be loaded before mounting a RAID pool

I have an mdadm raid pool mounted at boot in /etc/fstab, and I used the attribute nofail to not block boot in case of mount failure.

In my case I have Docker services with Volume Mounts to locations in my mdadm RAID pool, and when Docker started at boot it created a set of folders for every volume mount file or folder. Then some Docker services automatically started and wrote a bunch of default files within some of those locations.

This might be your problem if you run

sudo mdadm --assemble --scan

mdadm: Can't open /sys/module/md_mod/parameters/legacy_async_del_gendisk
mdadm: init md module parameters fail

Confirm diagnosis

lsblk should be missing any mention of md### UUID-from-/etc/fstab and have a blank value for MOUNTPOINT, even though sda/sdb/sdc/sdd/... devices are present

lsblk -o NAME,UUID,MOUNTPOINT,LABEL --fs

Preparation

Before solving the problem, you must make sure to clear out any files/folders created at your mdadm mountpoint that you have set in /etc/fstab

Stop docker if it created files/folders inside the mountpoint sudo systemctl stop docker
Feel free to back up the files with sudo cp -a /mnt/raidpoolname ~/justincase
Then delete contents with sudo find /mnt/raidpoolname -mindepth 1 -delete

Load RAID Pool Manually

sudo modprobe md_mod
sudo mdadm --assemble --scan
sudo mount -a

You should then be able to see your files at /mnt/raidpoolname

Restore mdadm mount at boot

mdadm 4.5 needs md_mod mounted before attempting to mount in my case, so using sudo edit /etc/mkinitcpio.conf and add md_mod to MODULES eg MODULES=(amdgpu md_mod) followed by a regeneration of initramfs with sudo mkinitcpio -P

Extra steps

That should be enough, but just to be safe you can take additional steps to avoid similar problems in the future

1. Make sure md_mod is required by mdadm, regardless of your /etc/mkinitcpio.conf
   echo "md_mod" | sudo tee /etc/modules-load.d/md_mod.conf

2. Make sure docker does not start without your raid pool being successfully mounted first
   sudo systemctl edit docker.service
   read the instructions and add the following in the space between the comments at the top of the systemd service /etc/systemd/system/docker.service.d/override.conf

[Unit]
Requires=mnt-raidpoolname.mount
After=mnt-raidpoolname.mount

Final steps

Restart docker if necessary sudo systemctl restart docker
Reboot the computer to make sure mdadm can load at boot

---

Note: This issue was directly addressed in mdadm pull request md-raid-utilities/mdadm#228 and was committed on January 13.

The fix is now available as package version 4.5-2 (Arch discussion) currently in the Arch Core-Testing repo