You are not logged in.
Thanks for the feedback, I'll dig through that.
Anyhow, I am a step further today when I was yesterday. New questions arise now, new post to be written. I will close this one.
BTW: installation protocol in my understanding is a listing of commands and outputs of the system as you get on F4, just all logged into a file for later analysis.
Installed twice before - once with User as admin and root - without automatic login and arrested at tty and installed again User with Sudo as Root
The normal Debian/Devuan way is to establish a root account, and a normal user account. Assign a root password during installation different from the user password. Then you work normally as user, and do administrative stuff as root in a root terminal. This way you do not need this sudo stuff.
Neither time have I been able to control change add etc involving 2nd and 3rd users ( which i HAD intended to use separately for each desktop).
Even with a cli installation without any desktop, the command line tools for adding/modyfing users and groups will be present (see Delgados response). Downside: you need to log-in in a root terminal and do the work by typing.
Depending on the desktop you will find one or the other graphical tool, e.g. Cinnamon brings a limited user/group graphical tool.
Gnome-system-tools is the more convienent tool, I also add this tool to my installations sort of at the end of the fine tuning.
BTW: For a multiuser-installation I would go away from slim (which isn't maintained any more anyhow) and use lightdm instead. Or sddm for KDE, which isn't my personal preference.
And it is not recognizing my data drives which have ALWAYS been instantly recognized and accessible no matter who is logged in, what desktop.
The desktop permissions are typically set by policykit. Each desktop should come with some set of rules to allow for user access to removable drives in combination with group permissions.
I tried twice now to install a Daedalus encrypted w/o separate /boot in a luks container holding a LVM for / and swap. Both legacy as efi installs fail from the 501 netinstall when writing the grub. Both in a Virtualbox VM.
With some poking around I managed to open a console with messages, I guess it was right CTRL F4. These messages do not really tell whats wrong, at least I can't see the reason for failure. I'd like to understand what is going wrong.
As far as I remember somebody told me that there also is some sort of protocol which I possibly could save. How do I get there?
Normal process, new stuff goes into sid at first, then eventually in testing, later into proposed-updates or backports.
239 is not in backports but in proposed-updates!!!
Yes, I have seen a similar topic in the German Debian Forum. There seems to be a new version 239 of the 470 driver package dripping down from sid. As far as I understood the discussion the new package is in proposed updates.
Thats why I am running Excalibur in VMs. Will keep that for a while ...
Looks like I wasn't patient enough. Running another update this evening fixed the issue. python3-apt has been upgraded in both VMs now, no more hold back packages.
Since some weeks I am getting the info that python3-apt is hold back from upgrading. I was waiting for a solution over time which did not happen.
# apt-mark unhold python3-apt tells me that the hold mark already has been removed.
Got two encrypted VMs, installed Dec last year, one with Mate, one with Cinnamon. Both show the same problem, both are up-to-date.
What can I do to solve the issue?
username ALL=(root) NOPASSWD: /usr/bin/setpci
Don't think you will have many issues. SATA2 or SATA3 does not really matter, CPUs are of same brand and similar generation. The only issue could be graphics, especially NVidea. But if you are running Chimaera or Daedalus, the old stuff should be supported as far as integrated graphics is concerned.
You should be able to plug in the drive into the new MB and it should work.
@fsmithred: never tried this yet because I wanted to used luks2. Is the grub as shipped with Daedalus now able to decipher luks2?
Yes, the traditional setup requires an unencrypted /boot partition. I use that setup. In the encrypted partition I do setup a LVM with 2 logical devices for / and swap. /home always is on a separate device, except for in a VM.
There is no need for special "early short command line detour for device mapping". You go through the installer step by step, and everything thats needed is prepared, no matter if legacy or efi installs.
With Arch Linux I could get away without that, but that requires a grub that can decipher luks2 and I think some tweaks inside of systemd.
Just my two cents: FDE works fine when pre-partitioning with a GParted Live and manual partitioning in the installer. Valid all along from ASCII to Daedalus.
You might consider that Chimaera even is too old for that laptop. Chimaera is oldstable anyhow, Daedalus is stable.
Just to make you fully aware: Devuan is nothing but a Debian deriviate. It makes sure you have init freedom, i.e. you can simply use the same kind of routines that did exist before Jessie and systemd introduction, and that you are NOT forced to use systemd. Inits supported like sysvinit, openrc, runit, ... All the rest of the SW is Debian, like this usrmerge topic. Too much to convert back by the small Devuan team.
I am absolutely happy that I have found Devuan, and that I can use the same kind of procedures to encrypt and decrypt my installations as used back in Etch, Lenny, Squeeze, Wheezy.
Make sure you understand who is responsible for what ... you are wasting your and our time here.
Go to Debian an bitch at them ...
... and NOT a Devuan invention ...
No, running the usrmerge installation will become mandatory with Debian 13/Devuan 6. There is NO way to avoid this. Lots of fighting and bitching, but no help, its a Debian project decision, and they are pulling all roots for not merging now during development.
Running with Daedalus on my workstation, I have set up two VMs with Excalibur based on Daedalus cli installations about 6 months ago. My first step before changing the sources to Excalibur was to run usrmerge as it is recommended now in the documentation. And both of them work well.
Well, I am working with LightDM/Mate re. Cinnamon. I myself had some trouble getting ASCII to work with consolekit, the elogind stuff fixed that. And that is what I copied into my reply.
When I do an apt list on my Daedalus I get:
# apt list libpolkit*
Auflistung… Fertig
libpolkit-agent-1-0-dbgsym/stable 122-3devuan1 amd64
libpolkit-agent-1-0-dbgsym/stable 122-3devuan1 i386
libpolkit-agent-1-0/stable,now 122-3devuan1 amd64 [Installiert,automatisch]
libpolkit-agent-1-0/stable 122-3devuan1 i386
libpolkit-agent-1-dev/stable 122-3devuan1 amd64
libpolkit-agent-1-dev/stable 122-3devuan1 i386
libpolkit-gobject-1-0/stable,stable,now 122-3devuan1 all [Installiert,automatisch]
libpolkit-gobject-1-dev/stable,stable 122-3devuan1 all
libpolkit-gobject-consolekit-1-0-dbgsym/stable 122-3devuan1 amd64
libpolkit-gobject-consolekit-1-0-dbgsym/stable 122-3devuan1 i386
libpolkit-gobject-consolekit-1-0/stable 122-3devuan1 amd64
libpolkit-gobject-consolekit-1-0/stable 122-3devuan1 i386
libpolkit-gobject-consolekit-1-dev/stable 122-3devuan1 amd64
libpolkit-gobject-consolekit-1-dev/stable 122-3devuan1 i386
libpolkit-gobject-elogind-1-0-dbgsym/stable 122-3devuan1 amd64
libpolkit-gobject-elogind-1-0-dbgsym/stable 122-3devuan1 i386
libpolkit-gobject-elogind-1-0/stable,now 122-3devuan1 amd64 [Installiert,automatisch]
libpolkit-gobject-elogind-1-0/stable 122-3devuan1 i386
libpolkit-gobject-elogind-1-dev/stable 122-3devuan1 amd64
libpolkit-gobject-elogind-1-dev/stable 122-3devuan1 i386
libpolkit-qt5-1-1/stable,now 0.114.0-2 amd64 [Installiert,automatisch]
libpolkit-qt5-1-1/stable 0.114.0-2 i386
libpolkit-qt5-1-dev/stable 0.114.0-2 amd64
libpolkit-qt5-1-dev/stable 0.114.0-2 i386Looks like the backend stuff is gone, don't know when, and apt list *backend* confirmed that.
Give the first part a try, it should replace the consolekit stuff.
You may have more luck when installing
# apt install libpolkit-gobject-elogind-1-0 libpolkit-backend-elogind-1-0 This should remove the consolekit packages and install elogind stuff instead. And that should work.
The solution is to wait for the solution ..... Have you raised a bug report? It may be fixed in the future.
The solution will come some day.
Well, if you want a stable system with minimum hasle stay with Daedalus.
If you are ready for the unexpected, try Excalibur or Testing, and for the often broken, try Ceres or Unstable. Freshly updated/developed packages first land in Unstable, when they have sort of proven to generally work, they may be forwarded to Testing.
Testing some day becomes Stable, but currently this is about 2 years to go.
You only should use Testing if you are ready to accept uneven ground and if you are willing to contribute to this phase, or if you have a good reason to do so, i.e. you have so new hardware that its supported in Testing only. Something for the experienced user. And Unstable will be even rougher.
Just my two cents ...
Well apt policy did not show the 32 bit version. Give
# apt list libcrypt1 -a a try.
Normally the 32bit version cab be installed by:
# apt install libcrypt1:i386I guess you are on amd64, a 64-bit system. Your application is a 32 bit application that requires a 32 bit library. You will need to enable i386 as architecture, if you haven't done so far, and then to install the 32 bit version of libcrypt1. Same procedure as for wine32.
# dpkg --add-architecture i386
# apt update
# apt upgrade
# dpkg -print-foreign-architectures
# apt policy libcrypt1According to the help dialogue system-config-printer is an old RedHat development. Its a tool for cups.
I like it the way it is. I want it locked to avoid that any user can modify settings and then complain that the printer does not work. Its an admin tool.
I am not aware about any configuration option that allows to remove the root password. I am not familiar with MX Linux. Maybe they have used a policy to remove the need for a password entry? Have a look at /usr/share/polkit-1/actions (Debian path).