You are not logged in.
system complaining systemd-logind
There shouldn't be a systemd-logind process at all, elogind should be handling the user session.
Do you have the elogind package? Sounds like your upgrade went wrong, you should probably describe exactly what you did.
Launch x as root starts X windows ok
*Never* do that.
Does it have to be executable?
No, it does not.
Can we please see the full output of
# blkid
# update-initramfs -u -k all
cat /etc/{fstab,initramfs-tools/conf.d/resume}The "#" symbol indicates that the command should be run with root privileges.
This guide shows how to use runit-init as PID1 in Devuan, it is primarily aimed at Jesse from Distrowatch ![]()
First, instal the runit-init package:
# apt install runit-initAPT will ask for a specific confirmation for the operation because replacing the init system is far from trivial, type in the required authorisation to install the package.
Next, reboot into runit-init. Check it has worked with
cat /proc/1/commThe output should be runit.
That's it!
EDIT: if a display manager is not required then add the user to the input group to allow access to the keyboard &c:
# gpasswd -a $USER inputThere are several ways to set the default audio device for ALSA: https://wiki.archlinux.org/index.php/Ad … sound_card
I already tried altering and removing the file
/etc/initramfs-tools/conf.d/resumebut this didn't help.
Removing the file won't work, you should tell us how you altered it.
This is the correct content if you don't want to use the swap to hibernate:
RESUME=noneEDIT: and rebuild the initramfs after you've changed the file:
# update-initramfs -u -k allboth my ASCII and Beowulf are encrypted LVMs with unencrypted /boot partitions. I tried the UUID of the encrypted partition. When I select the new entry, it clearly states that vmlinuz isn't found. Don't know yet how to deal with this situation.
Ah, right. Unfortunately I don't know how to deal with encrypted partitions in a GRUB configuration file.
A bit of sleuthing suggests that this might work:
menuentry 'Devuan beowulf' {
search --fs-uuid boot_uuid --set=root
linux /vmlinuz-4.19.0-9-amd64 cryptdevice=UUID=beowulf_uuid:cryptroot root=/dev/mapper/cryptroot ro quiet
initrd /initrd.img-4.19.0-9-amd64
}^ Again replace beowuld_uuid with the UUID of the (encrypted) root partition and replace boot_uuid with the UUID of the (unencrypted) /boot partition and make sure that the names of the kernel & initramfs images in /boot are corrrect (my last suggestion relied on the symlinks in the root partition that always have the same name).
Added EFI/BOOT/ to the efi partition and copied grubx64.efi from Beowulf into there
Did you rename the grubx64.efi file to BOOTX64.EFI?
EDIT: some UEFI firmware implementations are so broken that they will only boot $ESP/EFI/Microsoft/Boot/bootmgfw.efi
Try adding a file to /boot/grub/custom.cfg in your ASCII system with this content:
menuentry 'Devuan beowulf' {
uuid=beowulf_uuid
search --fs-uuid $uuid --set=root
linux /vmlinuz root=UUID=$uuid ro quiet
initrd /initrd.img
}^ Replace beowulf_uuid with the actual UUID of the beowulf root partition.
I don't know where to look for log/error messages for that
# dmesg | grep firmwareHow strange, having the contrib & non-free components added to the deb-src line shouldn't render the packages in main unavailable.
And just FYI ceres doesn't have -security or -updates repositories.
@HOAS: Since I am on PURE EFI without CSM, there is no boot entry for Beowulf in my ASCII grub menu. os-prober does only find the Win7 boot manager on the Win7-SSD. The same on a different mainboard with X570 chipset. Arch and Beowulf are present on the same NVME, I only can start the second OS via the boot override. I have just seen that an Ubuntu wiki claims that os-prober should be able to find another OS on the PC - here in my environment it does not work in two instances.
The UEFI GRUB version will only provide a menu entry for another system if it is also installed in UEFI mode. The non-UEFI GRUB version will only provide a menu entry for another system if it is also installed in non-UEFI mode.
So are all of your systems installed in UEFI mode? Check for /sys/firmware/efi, that will only exist if the system is booted in UEFI mode.
@Head_on_a_Stick: I had added the ceres deb-src lines in sources.list and apt update ran. When I search, this is the output:
Reading package lists... Done E: Unable to find a source package for openrc
You should show the actual line that you used and also the output of the apt update command so that we could see if it was being read. I did build the package successfully from the ceres source in a Devuan beowulf live environment so it definitely works and you don't need to use any Debian sources.
But anyway that probably belongs in another thread, we've got diverted somewhat here.
I checked the pool and there is no openRC packages there for Ceres. From what I understands, Devuan repositories uses Debian repos for most other packages? Will you share the Ceres version of openRC?
I can't share a link but I've just managed to grab the source from ceres in a live version of beowulf so it does work.
Did you add the deb-src line for ceres to your sources and update the package database before using apt source?
Here's my full procedure for the backport (lines beginning with "#" should be run as root):
# apt install devscripts dpkg-dev
# tee -a /etc/apt/sources.list <<<'deb-src http://deb.devuan.org/merged ceres main'
# apt update
mkdir -p ~/builds/openrc && cd ~/builds/openrc
apt source openrc && cd openrc-0.42
dch --bpo # fill in changelog entry and save the file
# mk-build-deps --install --remove
debuild -us -ucThere should then be an openrc-0.42-1~bpo10+1_amd64.deb package in the ~/builds/openrc/ directory that can be installed safely in your beowulf system.
To clean up afterwards remove the unneeded build dependencies and build scripts:
# apt autoremove openrc-build-deps devscripts dpkg-devAnd also remove the ceres source repository:
# sed -i '/ceres/d' /etc/apt/sources.listEDIT: and finally update the package database again:
# apt updateNote that the Debian buster release notes advise to use
# apt full-upgradeto upgrade from stretch so that should probably also be used for ASCII.
Reference: https://www.debian.org/releases/stable/ … ng.en.html
EDIT: your system will be far less reliable if you've just added the beowulf sources and used apt-get upgrade.
Tags in dwm are not the same as workspaces: https://github.com/w0ng/wongdev.com/blo … kspaces.md
You should join the ArchLabs forums and ask Nate about yaxwm, he is an extraordinarily helpful person.
Debian has 0.42 version of OpenRC available. Can we use the sources from the Debian version of OpenRC and rebuild it?
That version is available from Devuan's ceres repositories so you can grab the source from there and rebuild it for a beowulf base, follow this Debian wiki guide:
https://wiki.debian.org/SimpleBackportCreation
I've just tried that in my Debian buster box and it builds fine.
I do actually have an old OpenRC version in my OBS repository but I can't remember my password and the reset facility isn't working at the moment, I'll edit this post with a link once I get into it and update the package.
I've switched to Alpine Linux for my daily driver — the musl libc base ensures a fundamental incompatibility with systemd that I'm sure will be appreciated by the community here ![]()
The init system is busybox as PID1 with OpenRC as the service manager (with s6 as an alternative option) but I've changed to /sbin/openrc-init for PID1 by specifying it as a kernel command line option:
init=/sbin/openrc-initAll I had to do to get it working was copy the supplied agetty init script to TTY-specific versions and enable them to start:
# ln -s /etc/init.d/agetty{,.tty1} # repeat for other TTYs as desired
# rc-update add agetty.tty1 # repeat for other TTYs as desiredThe startup is blisteringly fast, quicker than my Debian buster box (which is using systemd) and the resource usage is pleasingly minimal.
Drew DeVault is one of the Alpine developers and he created sway so I'm using that to dodge the X server bloat monster (Wayland ftw!) and I've installed and enabled elogind so that I can disable the setuid bit on the sway executable and run a true rootless desktop.
I've also managed to get mokutil added to their testing repositories so that I could enrol my own Secure Boot key (created with their sbsigntool utility) and use it to sign the Alpine kernel and start the system with Secure Boot enabled.
It's an awesome distribution and *very* easy to package for (the APKBUILDs are almost identical to Arch's PKGBUILDs), very highly recommended.
Sorry but I am failing to understand the exact nature of the problem here.
On the weekend I may try HOAS hint to create the /EFI/BOOT/ entry as a copy of the beowulf efi entry. Lets see if that allows me to select either ascii or beowulf.
But surely you can already select either ASCII or beowulf from the GRUB menu, why do you want separate NVRAM entries?
I am considering to replace the mainboard
Good luck finding one that doesn't have a buggy UEFI firmware implementation. Based on my experiences on various forums the vast majority of motherboards have barely functional UEFI firmware, the manufacturers are clearly farming out their solutions to the lowest bidders.
Copy grubx64.efi to /EFI/BOOT/BOOTX64.EFI (on the EFI system partition), that will be booted even without an NVRAM entry.
It will probably be located at /EFI/devuan/grubx64.efi, use the find command to locate it.
Alternatively, use the grub-install command with the --removable option, which copies the bootloader to that location automatically.
You might be able to get efibootmgr working by mounting efivarfs manually:
mount -t efivarfs efivarfs /sys/firmware/efi/efivarsIt is also possible to manually delete the NVRAM entries from /sys/firmware/efi/efivars/ but it might brick your box.
I was almost convinced that Gnome 3.x will not work without systemd. But, proven wrong!
OpenBSD has had a port of GNOME for a long time and it's even available for Alpine Linux now thanks to elogind.
gcolor2, the simple GTK color picker is no longer available
Install yad and use
yad --colorThe interface is almost identical.
I had to pull the Intel wifi firmware from the ASCII live CD image, as the firmware included on both the ASCII and Beowulf full and netinstall didn't work with my Intel 8265 / 8275 (rev 78) wifi.
The contents of both packages are identical:
~/Downloads$ for i in 201{6,9}/lib/firmware/*;do sha512sum $i;done
1ff797870d3733f34343bc39ee74473276bba57159670cf90247d00d001ef35788eac52d2bd0a4da06f09fc2d5c9dac554d52eb4eea30d7a3070b9f226d1ec30 2016/lib/firmware/i2400m-fw-usb-1.4.sbcf
ef5d7495013a8d7e34106e5b6cdffc2f41e445505fecaa7874c2e7157004f98198dad5099e2c5a159ab34a18f3a3fb8a6843d02a787322c1aa7f8bf0f34398d9 2016/lib/firmware/i2400m-fw-usb-1.5.sbcf
3f4805378ea58c9465b82a4086b729d172b55caa868bba4b45726fad5adf9eb70c36df251f9183c24f891bf7064be9f54195b44a296350209179c9aaa728d346 2016/lib/firmware/i6050-fw-usb-1.5.sbcf
1ff797870d3733f34343bc39ee74473276bba57159670cf90247d00d001ef35788eac52d2bd0a4da06f09fc2d5c9dac554d52eb4eea30d7a3070b9f226d1ec30 2019/lib/firmware/i2400m-fw-usb-1.4.sbcf
ef5d7495013a8d7e34106e5b6cdffc2f41e445505fecaa7874c2e7157004f98198dad5099e2c5a159ab34a18f3a3fb8a6843d02a787322c1aa7f8bf0f34398d9 2019/lib/firmware/i2400m-fw-usb-1.5.sbcf
3f4805378ea58c9465b82a4086b729d172b55caa868bba4b45726fad5adf9eb70c36df251f9183c24f891bf7064be9f54195b44a296350209179c9aaa728d346 2019/lib/firmware/i6050-fw-usb-1.5.sbcf
~/Downloads$I would prefer to not use PA, but Firefox is compiled to require it, so it has to be there...
The firefox-esr package in Devuan beowulf is compiled with --enable-alsa and does not require PulseAudio, check about:buildconfig if you don't believe me.
That PA problem has cropped up a few times recently, search the boards for the solutions.
Note that Devuan's beowulf installer will only set OpenRC as the service manager, sysvinit will still be used as PID1.
To use openrc-init as PID1 see https://dev1galaxy.org/viewtopic.php?id=2788
a way to get Gnome to use wayland?
First log out from the graphical session, login to a console and stop LightDM:
# service lightdm stopThen try
MOZ_ENABLE_WAYLAND=1 QT_QPA_PLATFORM=wayland XDG_SESSION_TYPE=wayland exec dbus-run-session gnome-sessionThe first two variables are to make Firefox & Qt applications run under Wayland, the third variable should force GNOME to start a Wayland session.
If it works then create a file at ~/.xsession containing that line and select the "default X session" from LightDM. Not sure if LightDM can actually launch a Wayland session though so you might have to disable LightDM entirely and switch to a console login instead.
To automatically launch a Wayland GNOME session after a console login to TTY1 add this line to the end of the file at ~/.profile:
[ $(tty) = /dev/tty1 ] && MOZ_ENABLE_WAYLAND=1 QT_QPA_PLATFORM=wayland XDG_SESSION_TYPE=wayland exec dbus-run-session gnome-session@OP: do it yourself if you're so impatient:
https://wiki.debian.org/SimpleBackportCreation
I suspect that the dependency chain is far too tangled to make this a reasonable option, hence the lack of a package in the buster-backports repositories.