RAID 1 split after upgrade

[KDE_Raid]

Edit
I may be encountering this issue. I have had need to set the dirty bit due to bad super block errors in the past. Was the black listing of ntfs3 mentioned previously an attempt to work around this problem?

---

@linux-aarhus

When you boot into the snapshot with Linux 6.12.90, I have understood that those lines is not present in /etc/mdadm.conf - is that correct?

I think and you can verify if - using Linux 6.12.90 - edit the file /etc/mdadm.conf to contain (at the bottom of the file) the lines

ARRAY /dev/md/imsm0 metadata=imsm UUID=1c8946ad:ff1e4ca1:a87d8e06:47580247
ARRAY /dev/md/RAID_0 container=/dev/md/imsm0 member=0 UUID=c850e72a:7fc6b853:c49abf3f:f226e972

Save the file and reboot using Linux 6.12.90

Adding:

ARRAY /dev/md/imsm0 metadata=imsm UUID=1c8946ad:ff1e4ca1:a87d8e06:47580247
ARRAY /dev/md/RAID_0 container=/dev/md/imsm0 member=0 UUID=c850e72a:7fc6b853:c49abf3f:f226e972

To /etc/mdadm.conf had no impact. In my prior post, I indicated that my /etc/mdadm.conf file is empty on both the snapshot and upgraded versions. The output shown comes from the mdadm -D --scan command.

---

@linux-aarhus

When you are using Linux 6.12.48 - what is the output from

cat /proc/mdstat

I got the /proc/mdstat from the old snapshot:

cat /proc/mdstat
Personalities : [raid1]
md126 : active (auto-read-only) raid1 sda[1] sdb[0]
      2930264064 blocks super external:/md127/0 [2/2] [UU]

md127 : inactive sdb[1](S) sda[0](S)
      5040 blocks super external:imsm

unused devices: <none>

---

@linux-aarhus

If this doesn’t get your RAID online, I am thinking that some back-ported security-fix has created a regression with kernel 6.12.48 and 6.12.90.

I am thinking you should sync the 6.18 kernel and reboot - before reboot - ensure the /etc/mdadm.conf contains the lines for configuring the disk array.

For the kernels, my only lower options are:

• 6.6.141-1
• 6.6.135_rt74-1
• 6.12.89_rt18-1

I do not have the version you indicated and see no way to add it. I tried with 6.6.141-1 including the mdadm.conf file change as requested and still see no difference.

---

@linux-aarhus

And this must be the automounted partition - which at the time you ran inxi was not mounted - is that correct assumption?

Incorrect. The automounted partition is md126p2. The p1 variant is for a small Windows partition that doesn’t get used and does not need to be mounted. Never looked into why it doesn’t mount as it is unused.

---

I agree that NTFS is not a great choice for a file system on Linux. At the time it was created, NTFS was the only viable option. I have run into other issues with it, some sectors may be going bad, but nothing this problematic. I will change this when I can, I don’t have drives standing by in which I an easily perform backups nor rebuild. My only other drive of comparable size runs over USB 2.0, it would take days to move 3 TB over that connection. I’ll buy new drives that are larger and will move the data in time. That will take time.

Sadly I can’t even fully use the old snapshot. The update has changed things within the home directory (caches / settings) so parts of the old snapshot are now broke too. I’d need to use a mix of snapshots until the RAID can be rebuilt, not ideal.

Thanks for your comments so far. If you have any other suggestions, I’m willing to try.