I made the mistake of not breaking the raid array before rebuilding my system. I still have access to the NVME with the ZimaOS installed but the array disks are jumbled and one of the drives may not be original.
The OS does not read any storage devices other than NVME and lsblk in terminal shows some named disks but are labelled with zero file sizes (nbd0 thru nbd7).
Any help on how to fix this mess?
ZimaOS version 1.6.0-beta2
Also, I dont need to retain any data. Wiping and a restarting is fine for me.
So I have figured this out myself as no one replied here and Zima support has not replied to my email request for over three days.
I spent many hours of searching for and testing a fix and getting AI to help overcome some hurdles.
Below are the steps I used to unbrick my drives:
To format over a ZimaOS RAID or storage pool in Ubuntu, you must remove the ZimaOS RAID configuration by wiping the metadata superblocks from each individual disk, and then format the raw drives using standard Ubuntu terminal tools. [1, 2]
Open your Ubuntu terminal.
Run the following command to list all block devices:
lsblk
Identify the names of the drives that were part of your ZimaOS RAID array (usually formatted as something like /dev/sda, /dev/sdb, etc.). [2, 3, 4]
If the array is still actively running or mounted, you need to unmount and stop it so you can write over the disks. [5, 6]
Unmount the RAID partition (replace /dev/md0 with your actual RAID device name, e.g., /dev/md127):
sudo umount /dev/md0
Stop the active RAID array:
sudo mdadm --stop /dev/md0
[5, 7, 8, 9, 10]
To completely remove the RAID formatting, you must zero out the superblocks on each underlying physical drive. [1, 11]
Wipe the superblock on the first drive:
sudo mdadm --zero-superblock /dev/sdX
(Replace /dev/sdX with your first drive identifier, like /dev/sda)
Repeat this command for all other drives that were part of the ZimaOS RAID array. [1]
Now that the RAID metadata is gone, you can format the drives as single, usable disks. [12]
Clear any old partition tables from the drives:
sudo wipefs -a /dev/sdX
Create a new GPT partition label on the drive:
sudo parted /dev/sdX mklabel gpt
Format the drive with a fresh ext4 filesystem:
sudo mkfs.ext4 /dev/sdX
[3, 13]
Repeat these steps for each physical drive if you want to use them independently. [14]
· 1970 M01 1
If you can tell me:
How many drives you are dealing with
What your final goal for these drives is (e.g., setting up a new standard Linux RAID, single storage volumes, or network sharing)
I can provide the exact, customized commands you need.
[2] https://www.claudiokuenzler.com
[5] https://delightlylinux.wordpress.com
[9] https://www.datarecoveryspecialists.co.uk
[10] https://bbs.archlinux.org
[11] https://medium.com
[13] https://community.zimaspace.com
[14] https://github.com
Information: You may need to update /etc/fstab.
Yes, editing /etc/fstab is critical to prevent system boot errors and to automate mounting the newly formatted drives. [1, 2, 3]
If the old RAID array was configured to mount automatically, Ubuntu will hang during boot because that array no longer exists.
Open the file in a terminal text editor:
sudo nano /etc/fstab
Look for any lines referencing your old RAID array (e.g., /dev/md0, /dev/md127, or a UUID matching the old ZimaOS pool).
Delete those lines, or add a # at the start of the lines to comment them out. [4, 5, 6, 7, 8]
To ensure your newly formatted disk mounts automatically every time Ubuntu boots, add its unique identifier to the file.
Find the new filesystem UUID for your formatted drive:
sudo blkid /dev/sdX
(Copy the alphanumeric string inside the UUID="..." quotation marks)
Create a folder where you want to access this drive (e.g., /mnt/storage):
sudo mkdir -p /mnt/storage
Reopen the fstab file (sudo nano /etc/fstab) and add this new line at the very bottom:
UUID=your-copied-uuid-here /mnt/storage ext4 defaults,nofail 0 2
(The nofail option ensures your system still boots safely even if the drive is unplugged)
Save and exit (Press Ctrl+O, Enter, then Ctrl+X). [9, 10, 11]
Always test your /etc/fstab configuration before restarting your computer. A typo can cause boot failures. [12, 13]
Run this command to simulate mounting everything listed in the file:
sudo mount -a
[14]
If the command runs without displaying any errors, your configuration is safe, and your drive is ready for use. [15]
If you want to proceed, tell me your preferred mount path (like /mnt/data or /media/backup) so I can generate the exact line to paste into your fstab file.
[5] https://discussion.fedoraproject.org
[11] https://pheonixsolutions.com
[12] https://learn.microsoft.com
[13] https://forums.rockylinux.org
mke2fs 1.47.2 (1-Jan-2025)Found a gpt partition table in /dev/sdaProceed anyway? (y,N)
This warning appears because /dev/sda currently has a partition table (gpt) on it, and formatting the raw device directly will erase everything. [1, 2]
You should press y and hit Enter to proceed, but only if you want to format the entire drive as one single filesystem without standard partitions.
Formatting a raw drive (like /dev/sda) works, but creating a standard partition (like /dev/sda1) first is safer. It prevents other operating systems from seeing the disk as “empty” or uninitialized later. [3, 4]
If you prefer to do it the standard way instead of pressing y, type N to cancel and run these commands:
Create the partition partition:
sudo parted /dev/sda mkpart primary ext4 0% 100%
Format the new partition instead of the raw disk:
sudo mkfs.ext4 /dev/sda1
[5]
If you hit an error, let me know which path you chose (formatting raw /dev/sda or creating partition /dev/sda1) so we can correctly update your /etc/fstab file next.
[3] https://www.handyrecovery.com