You are not logged in.
Pages: 1
Finally I took the decision to invoke pulseaudio in the file /etc/rc.local in the local mode.
I start the X-Server manually. That's why I have no problem with a running pulseaudio in normal mode, that should be stopped before aktivating pulseaudio in the system mode. It may be not recommended, but it runs. So we can close this theme.
@fsmithred
I wrote about my installations with policykit. When noticing the difficulty to access to synaptic, I just had installed Beowulf after the information on devuan's website, that it is the new stable distribution. The policykit-packages were installed automatically. So I wonder why I don't get a small window to write root's password.
I remember that some years ago I had a similar problem with a former debian installation. As I have written: Using sudo and opening a terminal, I get the access. But I would like to know by principle reasons why this happens.
@LarsH
In my working installation I added a line to load pulseaudio in the system-mode to /etc/rc.local (we work with init, so this is possible).
Because we don't use systemd there is of course the need to write "#autospawn=no". This ist done.
I wrote about my other devuan-installation that exists to test changes of my configuration. There I deleted pulseaudio.
The gui for alsamixer, using gtk, shows two levels, that are not to movable In my last post I described a program for alsa, where it seems that the different channels are ruled, but nothing happens. Maybe a problem of the hardware.
Thank you for your answer
hmj
I changed to my installation for tests. there I deleted pulseaudio and load some alsa packages.
My mainboard uses an Intel-HD chip. I have to select between duplex-stereo and multichannel-
HDMI. For my hardware, normal stereo, I use line-output stereo rear (green), front headphones.
Using QasMixer I chose card 1, HD-Audio Realtek. So I can change the different levels - but without effect.
Using gnome-mpv I get the information:
Playback was terminated abnormally. Reason: audio output initialization failed.
Trying a "apulse" before the command to invoke gnome-mpv: gnome-mpv chrashes when starting.
This is the reason why I asked whether there is the need to reconfigure packages.
In my working-installation I continue with pulseaudio.
@fsmithred
policykit-gnome is installed, version 0.105-7
policykit-1 has version 0.105-25+devuan8
@larsH
Invoking amixer I get:
Simple mixer control 'Master',0
Capabilities: pvolume pswitch pswitch-joined
Playback channels: Front Left - Front Right
Limits: Playback 0 - 65536
Mono:
Front Left: Playback 5621 [9%] [off]
Front Right: Playback 5621 [9%] [off]
Simple mixer control 'Capture',0
Capabilities: cvolume cswitch cswitch-joined
Capture channels: Front Left - Front Right
Limits: Capture 0 - 65536
Front Left: Capture 3547 [5%] [on]
Front Right: Capture 3547 [5%] [on]
After the test I installed volumeicon-alsa. I suppose that it is necessary to delete pulseaudio to make it work?
Thank you
hmj
After installing the beta-dvd of beowulf (today upgraded to stable) I noticed two problems:
1. I use XFCE. Invoking synaptic normally I get a window to write root's password, then synaptic is opened. Now: nothing. I solved the problem by creating a button in the panel, setting a "sudo" before "synaptic-pkexec" and using "execute in a terminal".
So my question ist not: how to get access as a normal user (as root there is no problem), but; why don't I get the usual window.
Testing the behavior I opened a teminal and wrote "synaptic-pkexec". I got these messages:
==== AUTHENTICATING FOR com.ubuntu.pkexec.synaptic ===
Authentication is required to run the Synaptic Package Manager
Authenticating as: <full name>,,, (>system user>j)
Password:
polkit-agent-helper-1: error response to PolicyKit daemon: GDBus.Error:org.freedesktop.PolicyKit1.Error.Failed: No session for cookie
==== AUTHENTICATION FAILED ===
Error executing command as another user: Not authorized
Using ascii this problem did not exist. Anyone can propose me a solution?
2. Using tor-browser pulseaudio gives no sound. Using pulseaudio in the system-mode tor-browser has sound, but this mode is not recommended.
Pulseaudio has a lot of dependencies, I deleted pulseaudio in a test-installation (I have 3 different devuan installations bootable) to use alsa instead, but I got no effect. I suppose some reconfiguring may be necessary. What is to do?
For the future I think of a combination of alsa and jack.
Thank you for your answers.
Meanwhile I have another computer with the same mainboard and the AMD CPU Ryzen 3 3200G. I changed the computer by other reasons than mainboard and CPU. The system is running fine. Very seldom it occurred that my usb-keyboard and mouse were lost.
In the thread i got a lot of advices. Most of them didn't concern the reasons of my problem. So I will give some information, what is needed to use the CPU.
1. No problem using lilo with UUID. Indeed: Using /dev/sdaX as root devices instead of UUID makes no problems.
2. The package firmware-linux-free is installed, it is not necessary to select firmware-nonfree.
3. Before putting the harddisk in the new computer, I changed my selection in /etc/initramfs-tools/conf.d.d/driver-policy from MODULES=dep to #MODULES=dep, then I updated the
initrd.img for one kernel - it grew from about 7 MB to 21 MB. I installed the packages firmware-amd-graphics and amd-microcode.
4. So I could boot using the big initrd. I have a system with 3 installations, that I can boot from the same menu. That means: the boot directory contains more initrd.imgs. Using the
small (MODULES=dep) the root device was not found. Using the big one I had nearly succeeded, because with a kernel version 4.19 the framebuffer-font (terminus 11x22) did not appear, my screen was not detected, I suppose a default mode of the X-server. Installing a kernel of version 5.4 made everything run.
5. After these steps there is no need to keep the big initrd.img Set /etc/initramfs-tools/conf.d.d/driver-policy to MODULES=dep, and rebuild the initrd.img.
@Head_on_a_Stick
CPU: AMD Ryzen 3 2200G
Sorry, I didn't save the link, it was an opinion of someone, who answered a question concerning the microcode. I read it today, and I ever clean everything when closing the browser.
And, concerning grub: I don't think that the problem is caused by lilo.
@ rolfie
firmware-linux-free has been installed, but not yet firmware-linux-nonfree - thank you, I will try it. And, of course, after booting the debian live-dvd, I will look at the packages in synaptic, where I suppose that something could be needed.
Synaptic is rather comfortable, so I like to use it. Of course, I also use apt, when there is a need. When the backports in sources-list are activated, normally I get the backport-versions in synaptic. In the two cases I've written above I found the information by clicking on "versions", that synaptic in that special case didn't offer to download the actual versions. But thank you for that hint.
Finally, when you are using a Ryzen 7, I'm convinced to succeed. In 2018 there seemed to be some errors with randomize-numbers, produced by the kernel. Of course I will add some packages before changing the devices into the new computer.
Hans-Martin
(hmj are my initials)
@rolfie
Thank you - a lot of questions
It may be useful to give some informations about my experiences with linux-systems.
I started my linux life in the late nineties with a SuSE 5.3. Some years later I changed to debian, about since one year I'm working with devuan. I'm using lilo from the very beginning. Before starting with linux I used OS/2 and MSDOS. With lilo it was possible to start MSDOS on a second harddisk. I installed linux on a lot of computers including computers of friends. As a result I have a great experience with multiboot-systems.
- the HDD has a MSDOS partition-table. An early copy of the first 512 bytes is written in /boot under the filename boot.0800.
- I only used devuan-kernels 4,19 for devuan, debian kernel 5.4 for buster. As I've found, AMD ryzen 3 is supported since the version 4.17. Devuan stable has on the install-dvd 4.09. That's why the devuan live dvd is booting, but the amdgpu is not supported. But: it doesn't stop, the screen is not readable, that's all.
- only one hdd and one /dev/sr0 (read and write dvds). No encryption, no LVM, nothing special. Mouse and keyboard ar connected with USB 2..
- I put the devices back in the old PC, no problem at all. I write this on my old computer.
- Firmware: Searching in synaptic, i found an amd-microcode, made non-free by amd, and a firmware-amd-graphics. Software from the backports is chosen in sources-list, but, very funny, the stable-versions for these two packages are shown, I had to force the new backport-version to activate in in the list.
Concerning graphics: amdgpu.ko exists für 4.19 and 5.4. And I've found a posting, that the microcode could bring problems when booting.
ASUS-boards have extended possibilities to configure the bios. Choosing them you can exclude the UEFI and secure-boot possibilities. Some time ago I installed another computer with an ASUS-board and an AMD-cpu, where the same procedures in the bios made the system boot without any problem - but it was no ryzen-cpu.
memtest86+.bin is booting without problem, the kernels for linux are booting, but stopping.
- The program liloconfig will write a file lilo.conf. There you find the boot device by id and the partitions by uuid -> values: blkid and etc/fstab. I've found, some years ago, the modern systems using libata would need such a unique ID. There is another possibility to give a value to a partition by using PARTUUID= ... Should I try it? Until now I just use it for the swap-partition, file: /etc/initramfs-tools/conf.d/resume. And I could try to use /dev/sda1 and so on, because, when watching the informations given bei dmesg, I find sda1, sda2 etc., not the UUIDs.
My plan: I will boot the debian-live dvd, mount the partitions needed, using update-initramfs to renew the initrd.img.
By reasons that not concern the starting problems I actually change the computer against another with the same mainboard and the same cpu. So it will last some days before I can retry the installation. I'm sure: I'll find the right way ...
@rolfie
Thank you, but, as I wrote, i'm using lilo. And I changed some values of the new bios to eliminate UEFI. My old computer has no UEFI-Bios, so I changed the values of the new one to be able to start. It boots, but stops. I think with an activated UEFI-Bios I had no chance ti start the new computer.
@Dutch_Master
I'm using ext4. And I had verified after booting the Live-DVD, whether the UUID-names of the partitions existed. blkid and the content of fstab brought the correct values.
I'm actually using a computer with intel-cpu on a ASUS-Mainboard. Using lilo I change between Devuan ascii (backport kernel 4.19), beowulf and Debian buster (kernel 5.4). A very special buster without systemd, using slim. Everything runs without any problem.
I bought a new computer - ASUS-mainboard with AMD ryzen 3 - and put the harddisk in the new computer. After changing the UEFI-parameters in the BIOS I started.
The kernel was booted, but booting stopped with the message, that the root device with UUID ... was not found. But of course it exists. The problem is the same on ascii and buster. And my keyboard and mouse are not available, so the offer to use the programs of the initrd was not helpful.
Using a Debian live-DVD with buster and gnome everything is starting, the UUIDs of the partitions are found.
The mainboard has the possibility to connect mouse and keyboard using PS/2. Would searching adapters of the type usb to PS/2 be helpful?
Because the root device and the other partitions are not mounted, I have no entries in the files of /var/log.
Is the problem a lack of firmware, or do you have another idea? Is there a need to rebuild the initrd-images after starting the new system?
This is my first posting. So I give you the information, that I'm a german. If you like it you can send messages in german or french.
Thank you for your answers.
Pages: 1