jozzsi

> on systems without `mpathconf`, the multipath module in a non-hostonly initrd will be a noop most of the time. Perhaps `check()` should also call `mpathconf_installed()` in non-hostonly mode and...

> Arguably most users of non-hostonly initrds will not be multipath users. Requiring them all to either wait 10s or pass `multipath=off` on the kernel command line would be another...

One of the remaining failure is due to this diff for `systemd-tmpfiles-setup.service` ``` < After=local-fs.target systemd-sysusers.service systemd-journald.service --- > After=-.mount systemd-sysusers.service systemd-journald.service ``` Dracut unfortunately trying to edit `systemd-tmpfiles-setup.service` instead...

The remaining regression failure in TEST-41 with systemd v258 is intermittent - I nether have the skills, nor the time to look into this further. To reproduce `test/test.sh arch 41`

The failure seems to be related to plymouth service ``` rm -rf /usr/lib/systemd/system/plymouth-start.service ``` makes the test pass - suggesting that the issue is only reproducible when plymouth is available.

@bdrung What do you think of this approach ? Could you please help to conform and evaluate that this PR resolves the "portability policy" for Debian ?

closing due to lack of feedback.

@bdrung I think https://github.com/bdrung/dracut-ng/commit/0548f45c7e7ad64511455dc3c16efb3bdc4372f6 is a proper upstream solution as well, as I think the two concepts of hostonly-cmdline and sysroot are conflicting and hostonly-cmdline is not meant to be...

For reference: https://github.com/gentoo/gentoo/blob/master/sys-kernel/dracut/files/dracut-108-disable-ukify-magic.patch

> typo `distrobutions`. Fixed the typo. Thanks !