Hey. I'm a bit disturbed by seeing such a regression added so shortly before release. AFAIU the NEWS entry: >* If the /usr filesystem is on a RAID device and the INITRDSTART > setting in /etc/default/mdadm is not 'all', you will need to change > it to include that device. the value (which defaults to "none") needs now to be changed on all systems whose /usr is on a MD device. But so far, this situation (at least mine, where *everything* is on one single MD device) has been detected just perfectly well. I understand that one tries to make "more exotic" setups (e.g. /root not on MD but /usr on it) working,... in fact I've already wrote https://wiki.debian.org/AdvancedStartupShutdownWithMultilayeredBlockDevices years ago about the whole unfortunate situation in Debian. But the above seems to go just one further step away from things-working-out-of-the-box™ and making another place mandatory to have boot device information configured. As said, IMHO that really quite unfortunate, especially as we get more and more locations where people need to configure their stuff - which shouldn't be the case. Rather, the devices should be configured in their respective canonical configuration files/packages and systems on top of these (e.g. initramfs-tools or any initramfs-scripts/hooks (like cryptsetup)) should just detect things from there. Cheers, Chris.
Attachment:
smime.p7s
Description: S/MIME cryptographic signature