sys-kernel/dracut does not start plymouth in initrd if systemd module is enabled. (Everything works fine if systemd module is disabled) Plymouth is only started late in the boot after rootfs is mounted and main system takes over. [ebuild R ] sys-apps/systemd-208-r2:0/1 USE="acl cryptsetup doc filecaps firmware-loader gudev introspection kmod lzma pam (policykit) python tcpd -audit -gcrypt -http -qrcode (-selinux) {-test} -vanilla -xattr" ABI_X86="(64) (-32) (-x32)" PYTHON_SINGLE_TARGET="python2_7" PYTHON_TARGETS="python2_7" 0 kB [ebuild R ~] sys-kernel/dracut-036-r4 USE="systemd -debug (-selinux)" 0 kB Additional info: There might be some useful info in rdsosreport.txt of bug#504018. There is an error message: Failed at step EXEC spawning /usr/sbin/plymouthd: No such file or directory
This reminds me bug 499388. Could you test plymouth-0.8.8-r5?
Created attachment 372850 [details, diff] dracut-036-r4.ebuild.patch
Quick testing showed that plymouth starts early if plymouth-0.8.8-r5 is installed but I had some strange problems which resulted in a stuck boot (says /home can't be mounted even though it is actually mounted) if systemd module is enabled, so I will have to investigate further later this evening. Probably a different bug yet again...
Shouldn't this bug be marked as fixed? So with plymouth-0.8.5-r5 dracut starts plymouth from initrd. However, boot later stalls when real system has to start plymouth service. This only happens with plymouth-0.8.8-r5 if dracut systemd module is enabled. I'll submit a new bug report for this because this seems like different problem.
A blocker should be added anyway (comment 2).
(In reply to Alexander Tsoy from comment #5) > A blocker should be added anyway (comment 2). Dracut no longer rdepends on plymouth. It's a suggested dependency. I've suggested newer version of plymouth in postinst elog. + 24 Apr 2014; Amadeusz Żołnowski <aidecoe@gentoo.org> dracut-036-r4.ebuild, + dracut-037.ebuild: + Suggest newer version of plymouth, rels bug #504846. +