Issue: SW RAID1 Not Detected After Leapp Upgrade to AlmaLinux 9 (CentOS7 → Alma8 → Alma9 Migration)

[svitta]

Hi all,

I’m currently testing migrations from CentOS 7 to AlmaLinux 9 on a set of appliances that use software RAID1 (mdadm), and I’ve encountered a consistent issue during the final stage of the migration. Here’s the workflow and problem:
CentOS 7 → AlmaLinux 8 → AlmaLinux 9

What works:

• From CentOS 7, I initiate the leapp upgrade to AlmaLinux 8.
• The Elevate-Upgrade initramfs is generated referencing el8 packages.
• On reboot, the system boots into the Elevate-Upgrade GRUB entry and upgrades successfully to AlmaLinux 8.

Then from AlmaLinux 8:

• I run another leapp upgrade to move to AlmaLinux 9.
• This time, the initramfs is built using el9 packages.
• On reboot, the system tries to enter the Elevate-Upgrade environment — but fails to detect the software RAID1 device.
• As a result, no partitions are found and the boot drops into the emergency shell.

Additional Testing:

• To confirm this isn’t specific to our appliances, I set up a VM with software RAID1 and saw the exact same behavior:Upgrade-Elevateinitramfs fails to detect the RAID device during the upgrade boot.

Key Observations:

1. After the leapp upgrade, as it generates a /boot/initramfs-upgrade.img. lsinitrd /boot/initramfs-upgrade.img | grep mdadm doesn’t show that it has the mdadm.conf file
2. I tried to inject the mdadm.conf (i.e unpack the img file, insert the mdadm.conf file and repack it), but on reboot it shows dracut timeout errors

Important Note:

• Migration works without issues on systems using hardware RAID

Question:

• Has anyone else encountered this issue with SW RAID1 and Leapp?
• Is there any known workaround (e.g., including specific RAID modules in the initramfs, tweaking dracut, etc.) to ensure SW RAID is properly recognized during the el9 upgrade stage?

Appreciate any guidance or insights.
Thanks!