You are not logged in.
It may be worth to note, that all Devuan installer ISOs still come with all firmware in the ISO pool, for installation.
In my case . Daedalus netinstall iso why didnt contain gpu firmware ?
And what is the ISO Pool ?
Well .. i regret it for marking it solved. I mean i solve the issue by purging luit but a possible conflict lies there untracked and unnoticed..
Daedalus hasn't been officially released yet,& I think I read that that is the reason, as they haven't yet been set up....
And how a current Daedalus systtem gets security upgrades ?
And shouldnt Daedalus installer offer more sync info on the subject on the Package management stage ?
Indeed. Checked apt log.
Start-Date: 2023-05-02 11:37:07
Commandline: apt-get -o APT::Status-Fd=4 -o APT::Keep-Fds::=5 -o APT::Keep-Fds::=6 -q -y -o APT::Install-Recommends=true -o APT::Get::AutomaticRemove=true -o Acquire::Retries=3 install tas
k-console-productivity-
Requested-By: chomwitt (1000)
....sudo:amd64 (1.9.13p3-1)...
Things got more strange. I decided to run tasksel and remove all 'Console productivity' packages.
That happened indeed. But now sudo has been wiped out from my system!!
Invenstigating further....
I wonder what tasksel would have done during installation if i had selected both 'console productivity' and 'desktop enviroment'...
New installation of Daedalus.
In the Software selection stage : I selected Console Productivity (a choice done from curiosity) and no destkop.
After Finishing, reboot and login :
$ sudo apt-get install xorg
luit breaks x11-utils (<7.7.16) but 7.7..15 is to be installed
During a new installation of Daedalus ..
in the stage: Configure the package manager
in the step : Select security and release updates
I get the error :
Cannot access repository .The repository of deb.devuan.org could not be accessed. Its updates
will not be made available to you .
Turing to an Ctrl-Alt-f4 console i also see:
in-target: Ign: http://deb.devuan.org/merged daedalus security inRelease
-//- : Err: - // - Release
-//- : 404 Not found ip : 2a02:2a38:1:400:422a:422a:422a:422a 80
E: The repository does not have a release file.
I continued the installation . But all the related entries in sources.list were commented.
OK. i managed to install Daedalus.
During installation i got error of not being able to access a repo for updates.
Also after installatiion i had to add in sources.list non-free-firmware to install my gpu firmware. Why that change ?
Thanks @fsmithred and @ralph.ronnquist . Indeed using vga=798 (1600x1200 , 16bit) the installer proceeds normal. The gpu is 6950xt . I had already problems using it at Chimaera , and it was suggested to me to try Daedalus. So i guess it's not my 21:9 monitor to blame.
@pcalvert . I didnt find text option. (default install , expert and speech)
I used devuan_daedalus_5.0.preview-20230425_amd64_netinstall.iso
in my PC with a newly installed gpu 6950xt (rdna2) .
After i choose either default install or expert install i see a distorted image of
the installer that cant be read to proceed.
I found a 5year old thread of the same issue (i guess that devuan uses debian installer) :
https://forums.debian.net/viewtopic.php?t=132200
The image linked in that discussion is like what i see.
I tried the devuan_daedalus_5.0.preview_20230209_amd64_desktop-live image and it booted to gui with no problem. I write this line from using it.
In my Chimaera system i installed a new gpu amd 6950xt . And according to the wikipedia it's rdna2 navi21 architecture codenamed Sienna Cichlid.
My Devuan PC boots but in a smaller resolution and dmesg doesnt contain any related message. Also no gpu driver seems loaded ( lspci )
My initial setup of related install .debs :
xserver-xorg-video-amdgpu 19.1.0-2 installed but it doenst have new version backported .
firmware-amd-graphics 20210315-3 But there is a new version of the firmware backported.
So i installed :
firmware-amd-graphics 20230210-4~bpo11+1
In the firmware file it contains related firmware:
amdgpu/sienna_cichlid_ce.bin
But the issue remains.
I remember a Phoronix article on amdgpu changing the way it releases the firmware. In that case could it be that a newer version of amdgpu is also needed?
Also it seem to me strange that both versions of xserver-xorg-video-amdgpu support the same chip families. I think there is something no rigth here. What is the relation of the driver and the firmware .debs? Can one support a family and not the other ?
Package: xserver-xorg-video-amdgpu
Version: 19.1.0-2
This package provides the 'amdgpu' driver for the AMD Radeon cards. The
following chip families should be supported: Bonaire, Hawaii, Kaveri, Kabini
Mullins, Iceland, Tonga, Carrizo, Fiji, Stoney.Package: xserver-xorg-video-amdgpu
Version: 23.0.0-1
This package provides the 'amdgpu' driver for the AMD Radeon cards. The
following chip families should be supported: Bonaire, Hawaii, Kaveri, Kabini
Mullins, Iceland, Tonga, Carrizo, Fiji, Stoney.
Refs:
Xorg RadeonFeature althougth not sure if it applies to amdgpu .
Archwiki page on amdgpu . Althought archwiki is another distro sometimes you find usefull info. not in my current issue thought..
Devuan GNU/Linux 5 (daedalus/ceres)
After doing :
$ sudo apt-get update
$ sudo apt-get upgrade
$ virt-manager --debug
--if i try to 'open' the window of a virtual machine virt-manager
crashes with the following message:
(virt-manager:19736): libsoup-ERROR **: 08:54:50.904:
libsoup3 symbols detected. Using libsoup2 and libsoup3
in the same process is not supported.
$ sudo apt-cache policy virt-manager
virt-manager:
Installed: 1:4.1.0-1
Candidate: 1:4.1.0-1
how can i correct this?
I've checked https://tracker.debian.org/pkg/virt-manager
and bug reports upstream at the project's bug tracker
but i havent found something.
But in searching the interent i found arch related opened issues FS#75823 where the cause is said to be libgovirt on which many applications depend
I've managed to connect to a running vm from the command line:
$ sudo virsh list --all
Id Name State
-----------------------------------
6 Hyperbola_min_vm running
- Devuan_vm shut off
$ virt-viewer --connect qemu:///system 6
So the issue in my case dont seem related to virt-viewer
I thing we live in an era that foss land is attacked. Not exactly the whole word . One letter is attacked.
For me the attack was formely initiated the day RMS visited Redmond (what was on his mind??...) .
Days letter the web lynching of RMS started. Paidophile, autistic , dirty bearfoot hippy..etc.
The great irony?. The MIT scandal raised serious ethical questions about funding from malicious and unethical actors to institutions
that we prefer to envision as public and motivated by higher ideas. Education, knowledge etc.
But the second greatest irony? Many Institutions in the foss land (SFC,Redhad,FSFE) aligned with MIT's heads scapegoating agenda saw that as an opportunity
not to discuss the matter of ethical donatiions but disgracing publicly RMS persona announcing to the plubs the new inclusive order of things.
In that new era.. Linus and Richard are in the social dysfunctional spectrum and in foss fora the new ideas of sponsor hunting and corporate friendliness have taken the sky..
HUGE piles of money (huge propably in our eyes for a big corp propably one medium scale strategy campaing.. for those uninitiated a study of how Uber conducts those affairs should be a good start up tutorial...) and donations from SFC to OpenBSD and Gnome have been softening the ground to let the new order bloosom.
Foss Newsites call us to be thankfull to our benovalent new masters in the fossland.
I've been familiarizing myself with debtags and the CLI command 'debtags'.
I found it helpfull and I wanted to ask if a GUI utility with debtags support , sth like packagesearch exist in devuan.
Thank you.
I've created an installation video . I'll use that hint in future versions.
Installing Devuan Chimaera's netinstall iso in a qemu vm i see the initial installation menu and choosed Help.
There by pressing F keys i can see various help menus but i cant find a way to return to
the main menu.
Also i tried to hit Enter to install but i get:
No DEFAULT or UI configuration directive found!
Interestingly invocation section of man page of bash dont refer to profile.d .
So maybe it started as a distro practice that stuck and now it's used a a global config-set up space ?
I found an interesting discussion here.
It seems that profile.d is used in some distros that use some shells that parse that folder for application specifics setups.
In my devuan installation:
/etc/profile.d$ ls
bash_completion.sh flatpak.sh gawk.sh vte.csh gawk.csh vte-2.91.sh
But that wont work in a case like mine but only if libvirtd was reading some shell enviroment
variable to determine what net setup to enable when started.
So i guess profile.d is one way to setup an 'enviroment' for applications that would look to certain env variables.
If an app looks only to a config file profile.d it's not needed.
In my case i found first here about a command that fixes
a network error in virt-manager. Now as @xinomilo told me that can be set either with a terminal command or a gui setting that changes a relative setting.
I propably was unlucky finding only the first 'solution' that make me understand that a command fixes an error . But now it seems it's just an application
setting that propably sitting in a config file and there are two ways to change that 'file' and propably to edit it by a text editor.
And maybe virt-manager shouldnt either report it as an error! That small detail could have put me into a different fix mentality and not in the 'config' mentality.
To sum it up. It was a setting issue masqueraded as a pseudo 'correct functioning' issue.
chomwitt wrote:From where ? From general options or a VM options ? I didnt find anything.
Connection Options.
if you already connected virt-manager to local libvirtd : choose that (qemu:///system) connection and :
(menu) Edit -> Connection Details -> Virtual Networks tab -> choose network, and check "Autostart On Boot" on network Details.
I found it. Thanks!
Reading here https://www.linuxfromscratch.org/blfs/v … ofile.html
it seems that profile.d is for scripts that set up the shell enviroments.
So if i wanted to setup an application and it's initial set up i should do it from somewhere else.
But i've just read an askubuntu post where network config is attempted from a profile.d script.
.. (you can also change this from virt-manager/GUI).
From where ? From general options or a VM options ? I didnt find anything.
Head_on_a_Stick wrote:Was the script executable?
Sorry, that was a stupid question. Files in /etc/profile.d/ are sourced rather than executed. D'oh!
That explains why other scripts in profile.d werent executable.
Generaly i realized that profile.d is for preparing the system wide enviroment (of bash only or more generally ?)
So a more sound approach would be to start libvirtd and then make sure the default networking is enabled .
virsh is a libvirt client, so yes, libvirtd must be started in order for virsh to work..
although this https://dev1galaxy.org/viewtopic.php?pid=36441#p36441 makes your command unnecessary.. (you can also change this from virt-manager/GUI).
Yes i understand that. But i was curious to see if indeed a 'root' command can be executed in profile.d
So i could make the script wait and poll the processes and when libvirt is executed then to execute the command!
I made it executable but still nothing.
#!/bin/sh
export TESTPROFILEDOTD=1970
virsh net-start default &
The test var is changed but not the net default state.
I suspect that virt-manager and libvirtd process must have started first.