You are not logged in.
apt update, apt upgrade
Did you run apt dist-upgrade too?
I'm currently installing excalibur on my new desktop, which is not as much fun as I hoped.
The proprietary nvidia driver seems not like kernel 6.10. Beside that and some uefi boot BS, everything seems to work. But I don't use/have KDE.
EDIT:
Debian's testing netinstall.iso throwed errors after booting, so I ran bookworm and daedalus isos and migrated them to excalibur.
I don't know synaptic, but can parse the list to apt. If you hate terminals, just ignore this.
"listapt.txt" contains _all_ debian packages in the installed repositories (--installed or --manual-installed would do a selection like below).
You want to re-install the ones marked with "[installed]", not "[installed,automatic]". Have a good look at "[installed,local]", which were either downloaded or not in the current repositories.
cat listapt.txt | grep '\[installed,local\]'
Let's prepare the installation:
There will be errors; work on cleaned.list.txt, until apt succeeds. Then omit "-s" in the apt command.
cat listapt.txt | grep '\[installed\]' | awk -F/ '{print $1}' >cleaned.list.txt
apt install -s $(cat cleaned.list.txt | awk '{printf"%s ", $0}')
Too many errors? Maybe just the tasks for the big picture?
apt install $(cat cleaned.list.txt | grep ^task | awk '{printf"%s ", $0}')
Hope this helps, and yes I'm about to do that myself.
Interesting, didn't know that cp has a device mode. Fist thought was you ment cat, which equally works here.
Today I looked at some back-up copies in /root/_backup/usr/local/bin/ ... and found symliks "corrected" too.
That's not funny!
apt purge symlinks
Thanks fro the reply. The accident happened about three weeks ago. usrmerge was installed months ago, before switching to excalibur.
After more digging, I guess it was the package symlinks, or a "cleaning" job using it.
All symbolic links in /usr/local/bin (pointing to another directory) have the same updated timestamp. In my case "2024-07-01 19:25:17", or 20 seconds after installation of the package "symlinks".
# /var/log/dpkg.log
(...)
2024-07-01 19:16:23 startup packages configure
2024-07-01 19:24:58 startup archives unpack
2024-07-01 19:24:58 install symlinks:amd64 <none> 1.4-4
2024-07-01 19:24:58 status half-installed symlinks:amd64 1.4-4
2024-07-01 19:24:58 status triggers-pending man-db:amd64 2.12.1-2
2024-07-01 19:24:58 status unpacked symlinks:amd64 1.4-4
2024-07-01 19:24:58 startup packages configure
2024-07-01 19:24:58 configure symlinks:amd64 1.4-4 <none>
2024-07-01 19:24:58 status unpacked symlinks:amd64 1.4-4
2024-07-01 19:24:58 status half-configured symlinks:amd64 1.4-4
2024-07-01 19:24:58 status installed symlinks:amd64 1.4-4
2024-07-01 19:24:58 trigproc man-db:amd64 2.12.1-2 <none>
2024-07-01 19:24:58 status half-configured man-db:amd64 2.12.1-2
2024-07-01 19:24:59 status installed man-db:amd64 2.12.1-2
2024-07-02 15:23:51 startup archives unpack
(...)
package: symlinks
summary: scan/change symbolic links
version: 1.4-4
description: Symlinks scans directories for symbolic links and lists them on stdout.
Each link is prefixed with a classification of relative, absolute, dangling, messy, lengthy or other_fs. .
Symlinks can also convert absolute links (within the same filesystem)
to relative links and can delete messy and dangling links.
Hi,
a way to have programs from /opt in the $PATH is to make symbolic links in a bin directory.
One day (begin of July) all link targets in /usr/local/bin changed from absolut paths to relative paths.
Most likely during an upgrade. On excalibur, not on daedalus.
cd /usr/local/bin
ls -l kpackage
kpackage -> /opt/trinity/bin/kpackage # original
kpackage -> ../../../opt/trinity/bin/kpackage # now
Why?
As EDX-0 suggested in #2: Just create a swap file of decent size, and you should be good to go.
It is not too complicated and described in e.g.: https://wiki.archlinux.org/title/swap , section 3.1 "Swap file creation".
@fsmithred: Thanks for debsums!
@SpongeBOB: My personal preference is to upgrade - beside the good advices above.
Patience. excalibur is currently 'testing' and such issues are usually solved in one or some days.
Not sure about the architecture, but do you have daedalus-security in /etc/apt/sources.list?
EDIT: Ups, there is another thread respective this issue.
Ich kaufe ein "G" und möchte auflösen:
Google Chrome,
Vivaldi,
Opera
and worst of all f*ck Mickysoft's Edsh.
EDIT: And the best part of all existing Android "Apps".
It would be easy if i could just make that machine somehow forget that it is supposed to access ssd before booting from optical.
But the first attempts at that have not been successful.
Did you try to reset the BIOS/(U)EFI settings to system default?
There should be a menu somewhere saying something like "load save defaults".
Maybe missing firmware?
apt list --installed | grep firmware
firmware-intel-sound
firmware-realtek
firmware-linux
firmware-linux-free
firmware-linux-nonfree
Then maybe it is:
apt install linux-image-arm64 -s
The switch -s is for simulation - to see what would happen.
Package info:
https://pkginfo.devuan.org/cgi-bin/poli … 4&x=submit
Hi,
is the generic linux image installed?
apt list --installed | grep linux-image for a check and in case it's not:
apt install linux-image-xxxx
where xxxx is amd64 on my PC, but something else for the arm-CPU of a Raspi 3+.
History: When devuan forked debian, the std. network interface name was eth0 on both systems, so a migration does not change anything in this respect. And this is (probably) the case for Buster.
Some years later (release?) debian changed the naming convention to the current expression, devuan did not. Now the default net device names are different.
apt-get install net-tools works on any recent debian, no idea what rackspace did.
Not allowing java script by default can help.
I usually have the extensions "ublock origin" and "umatrix" running with firefox.
to a) Buster probably still used eth0/eth1 itself (don't remember, when the change was in debian).
Then there was no net-device-name change during migration.
to b) There is no network-auto-configuration tool involved during migration.
E.g. network-manager or connman would do such things, if installed.
When you use plain dhcp without such programs, it's your task to fix the network.
net-tools are not installed by default anymore, but present in the std. repos.
Regards
Thats fine, but i cant install usrmerge, i cant install anything now, apt/dpkg is broken.
EDIT: this was a reply to someone but they deleted the post.
Sorry for the inconvenience. The post's online time was less than 5 minutes. Tight timing!
It was the 3rd post in less than three minutes (after ralph and soren), saying that you need usrmeger for ceres.
BTW: In such cases, It is possible install a downloaded package via dpkg -i /path_to/some-package.deb manually.
Anyway. Good to hear that you made it!
Probably I confused you. That was not intended.
Both points in my 1st post are completely optional! And proposed, since the offical migration did not work in your case.
#3
Clearly after that reboot I'm ending up with eth0 and eth1
Which clearly indicates "net.ifnames=0"
#5
So first of all, adding net.ifnames=0 to the kernel parameters (and I did make sure that I actually saw that option in /proc/cmdline) did NOT cause it to use the enX0/enX1 interface names...they were still set to eth0/eth1 after rebooting.
Of couse not. "net.ifnames=1" would accomplish that. (At least on plain debian and without addional renaming).
I have no idea what "Rackspace's Debian Bookworm VM" might imply, but anyway.
Network:
The configuration is changing during migration. One option is to keep the configured "enp0s0"-alike debian device name:
$ su -
# echo 'GRUB_CMDLINE_LINUX="net.ifnames=1"' >>/etc/default/grub ### OR to 'GRUB_CMDLINE_LINUX_DEFAULT'
# update-grub
To further split migration in smaller parts, you could try to
# apt install sysvinit-core
as debian (reboot needed!) and then migrate to devuan.
Just fishing in the dark ... . You mentioned "HTTPS-enabled repos":
There is an issue with devuan repos and https - I don't remember details.
"deb http://" is recommended, while "deb https://" should not be used for *some* reason with the (merged) devuan repos.
cat /et/apt/sources.list | grep ceres
deb http://deb.devuan.org/merged ceres main contrib non-free non-free-firmware
Reminds me on a Raspi 2.
A real slow micro-computer, but it can be used for video playback, if you purchased a license for "video decoding on the graphics hardware" additionally.
Chrome OS may include proprietary firmware with such a feature?
Are you using pipewire audio?
There are usually 3 programs to start with the X session (pipewire, pipewire-pulse, wireplumber. There is a forum thread).
Maybe, these are auto-started on XFCE, but not on Ratpoison?
I have a clonezilla backup of Chimaera install that works perfectly.
In this situation, I would have upgrade chimaera to daedalus.
EDIT:
The one line above sounds stupid, sorry. I'll try again in detail:
There is no need to re-install. Doing a dist-upgrade with apt works fine. Benefits are: Manual configurations are preserved, less time consuming, everything is basically as it was. The process is not too complicated and in the worst case you can recover chimaera.
(edit /etc/apt/sources.list accordingly; apt update; apt upgrade; apt dist-upgrade and it's done)