I recently installed a fresh minimal copy of Fedora 34 using Anaconda and this fairly simple Kickstart file (see below). To save you some reading, the Kickstart file creates 3 software RAID1 arrays (/, /boot, and swap) and installs Fedora 34 on top. When Anaconda completes, I'm presented with a working system running on top of a bootable software RAID1. So long as I continue to use the initramfs that was created for me by the installer, I can unplug a drive (thus degrading the RAID1 array), and continue to still boot the system. But, as soon as I run dnf upgrade and pull in a new kernel (updating 5.11 to 5.13), the new initramfs that gets generated will lose the above functionality (booting off a degraded array). The system will still boot when all RAID1 members are present, but as soon as a single member of the RAID1 is lost, I get dropped into a dracut emergency shell and then need to manually turn all the inactive arrays active via a series of mdadm --run /dev/md[X] and mdadm --readwrite /dev/md[X] commands. Any idea why Anaconda+Kickstart can create a working initramfs that supports booting off of a degraded software RAID1 array, but dracut can't when it's automatically invoked by dnf update? For additional information see: Kickstart file: https://gist.github.com/zacharyfreeman70/ead4c8a9d2fec58753d7c295f39cfc00 GitHub issue: https://github.com/dracutdevs/dracut/issues/1593 Reddit post: https://old.reddit.com/r/Fedora/comments/pdqw3v/initramfs_built_incorrectly_after_dnf_upgrade/
