September 27, 2024
- PulseSecure no longer works on nb-dual (Ubuntu partition), so reinstall. Updated v22.7.R1 over 9.1.R11. Connected.
September 17, 2024
- Made some space by removing old snap versions with LANG=C snap list --all | awk '/disabled/{print "snap remove " $1 " --revision=" $3 }' | sh - .
- Removed some large files in my ~/Downloads after an inspection with du -aBM 2>/dev/null | sort -nr | head -n 50 | more
September 3, 2024
- Adobe Acrobat is no longer under my license (after the raised the price), so I couldn't edit a PDF anymore, or convert it to a word document.
- Checked UvA software page, and tried apps.uva.nl.
- Uploading files to the WebClient works, although I had some trouble to find the location of the uploaded files (and half of the OneDrive files are not visible).
- The location of the uploaded file can be found on PRD-R5-2 / Virtual Station / Uploads.
- The Reader on the webclient can also not edit or save-as.
-
- Tried to install okular on WSL 1, but libQt5Core.so.5 is not found (while it is present in /lib/x86_64-linux-gnu, and other libraries in this folder can be found. All logical links also seem to work, persmission also OK).
- Could execute evince assessment_.pdf, after setting export DISPLAY=0:0 (and running X-server). Yet, also there no editing.
- Easiest solution is to open the pdf in Microsoft Word.
August 21, 2024
- Continue with the HP Omen.
- The gdm fails on a missing gconf.
- Followed these instructions to install gconf-service.
- Looked at the ACPI error. With F10 going into BIOS, changed Graphics from Hybrid to Discrete. Still no gdm.
- TIme to install nvidia-drivers again.
- Installed according to this page.
- Tried again the reinstall of ubuntu-desktop from here, but still the gdm and other display managers fail.
- Tried a sudo do-release-upgrade -d, but that fails on installArchives().
August 19, 2024
- Got the apt update working again with the trick from askubuntu: adding [trusted=yes] to the repositories.
- Looked here, and made a link between resolvectl and systemd-resolve.
- Checked /etc/systemd/resolved.conf, but every service is commented out.
August 18, 2024
- Trying to recover Karlijn's Omen machine again.
- Starting Ubuntu directly was hanging on two errors. Bluetooth reported a Malformed MSFT vendor event: 0x02 and usci_acpi gave a PPM init fialed on USBC000:00:error -ETIMEOUT.
- Recovery mode worked, although sudo /lib/systemd/systemd-resolved didn't give me wireless internet connection in recovery-mode.
- Using wired ethernet I can reach the internet.
- Yet, sudo apt-get update fails on a failure on apt-key.
- Looked at the suggestions ubuntu forum, but the script doesn't remove any key (new print format of apt-key --keyring list.
- The list gives two old keyrings. Moving those two packages away (pgdg.list and mattermost.list) from /etc/apt/sources.list.d only partly helped, because the problem is in the archive.ubuntu.com jammy from /etc/apt/sources.list.
- Suggestion was to look up man apt-secure, but that fails. The packege ubuntu-keyrings is installed, the ubuntu-keyring-2018-archive.gpg can be found in /etc/apt/trusted.gpg.d/.
- Looked at the focal man page, but the option allow-downgrade-to-insecure=yes had no effect. No idea where to set the option Acquire::AllowInsecureRepositories. The Jammy man page gives the same information.
July 7, 2024
- With gdb3 active, the boot fails again.
- Luckily directly executing sudo /lib/systemd/systemd-resolved gives me internet connection in recovery-mode.
- Looked at the three methods to see what is installed of the nvidia drivers
- The command dpkg --get-selections | grep nvidia | grep install showed that there are several deinstalled packages, but that only libnvidia-compute-535:amd64 is currently installed. No drivers are installed with snap.
- This post indicated that nvidia blacklisted the nouveau-drivers, so tried echo 'nouveau' | sudo tee -a /etc/modules. Yet, I forget to remove the /etc/X11/xorg.conf. Try that, because GDM can still not start detect available GPUs. Yet, this file did not exist. Tried a reinstall the niveau drivers, but that still didn't help.
- I should have read the post further, and modified the blacklist-nouveau.conf and should have done update-initramfs.
- Followed the official instructions, and installed the latest nvidia drivers. There is an effect: GDM can detect the available GPUS, but still fails.
- Note that several services fail, also the CUPS scheduler, Bluetooth, Time and Date, Tunderbolt and docker services fail.
- If I do sudo apt update, I get many warnings due to the depreciated apt-key.
- Looked with sudo apt-key --keyring /etc/apt/trusted.gpg list and only two packages have keys (owncloud and PostgreSQL).
- Reinstalling gnome-shell didn't help, but CTRL+ALT+F2 gave me a terminal without the need to go in recovery mode.
- Got a good suggestion: look at the errors during boot with journalctl --boot --priority=err
- The kernel gives ACPI BIOS errors, and a nvidia-gpu i2c timout error.
- After that the systemd errors occur, which are permission denied errors. That is strange, because I can do it from the command-line. Seems more that systemd cannot create a directory for its logs.
- Looked at this post, and found that /lib was a redirect to /usr/lib, which had not permission 755 like /usr/lib32 and /usr/lib/lib64
- First trying a firmware update for the HP Omen 15-ek0120nd, it is version F.17 from 17 May 2024.
- Also Device Firmware is updates, such as the TitanRidge 2C (version 06 to 09)
- After rebooting, Windows also starts a lot of updates, including version 22H2.
- The Firmware update didn't help. Note that not all services fail, for instancse the snap service does start.
- With systemctl list-units you all services (162), with in red the five services that failed:
- bluetooth.service
- bolt.service
- systemd-hostnamed.service
- systemd-resolved.service
- systemd-timesyncd.service
July 6, 2024
- Karlijn's Omen machine had problems with booting.
- Problem was that the CUPS log was 76% of the storage (600 Gb).
- Received constant warnings that the CUPS service was not starting. At the end removed the printing service with sudo snap remove --purge cups.
- Removed the old 510 version of nvidia-driver, but this removed the whole desktop.
- Reinstall could be done with sudo apt install gdm3 ubuntu-desktop, but this fails because there is no internet-connection.
- Installed net-tools, but that required several libc6 dependencies (and sudo dpkg --configure -a.
- So, I can see with ifconfig the wired connection, I can do ping 8.8.8.8, but the NameServer fails to start up.
- Directly sudo /lib/systemd/systemd-resolved adds two dns-servers to my system (84.116.46.23 and 84.116.46.22). Now I can do ping www.google.nl and also do a sudo apt upgrade (which installs the pending signal-desktop package).
- Did sudo apt install gdm3 ubuntu-desktop followed by systemctl restart gdm, which gave again Failed to start CUPS Scheduler and Failed to start Gnome Display Manager.
May 31, 2024
- Had again problems with WSL. Killing in the Task-manager all WSL-processes, doing wsl.exe --shotdown in a PowerShell and restarting LxssManager in Services solved this issue without reboot.
May 17, 2024
- My XPS workstation has problems again.
- Recovery mode still works
- Seems that a install was halfway killed. Had to do a pkg recovery.
- Nvidia keyring was old, so updated with the method indicated in this post
- Could update many programs. The boot goes further, but has a flashing mounting disk.
- Back to recovery mode, several installs are held back, including cuda and cuda-drivers. Instaling them manually, which results that many cuda-12-4 packages are installed.
- Removed some old cuda-packages, including cuda-9.2
April 2, 2024
- Rebooted nb-ros, still no wired connection can be found. Last Thursday I tried the connection with nb-dual, which worked.
- The wired connection shows on its leds traffic, so it seems no hardware problem.
- Found this post, which suggested to do sudo lshw -C network, which only gave one logical name (wlan0)
- The old GUI nm-applet no longer worked, but nm-connection-editor gave the possibility to edit the connection. Seemed that the connection was only half saved, so included a netmask (255.255.255.0) and a Gateway. Device is still not visible with nmcli device.
- The command nmcli con shows that the connection is known, but not coupled to a device.
- There seems to be two problems. One is that with Ubuntu 18.04, they switched from one controller to another (netplan and NetworkManager).
- Looked at this suggestion, and modified /usr/lib/NetworkManager/conf.d/10-globally-managed-devices.conf to add a suggestion for type type:ethernet. Restarting the NetworkManager didn't not resolve it.
- Looked in the /etc/NetWorkManager/system-connections, but both 'Wired connections' were the same, except the settings which I just edited with nm-connection-editor
- Looked in /etc/network and added an eth0 entry in the file /etc/network/interfaces. As a result, the connection shows up (from type ethernet), but also not connected to a device.
- Strange enough, no leds led up when I use the other Dell adapter.
March 28, 2024
- Had the problem that when connected to the camera, I couldn't connect to internet (wired had preverence on wireless connection)
- This advice, but when I select Use this connection only for resources on its network I cannot apply it (I am in manual mode).
- Next followed the advice of this post, although I had to use the connection, not the device name.
- Looked up the connection name with nmcli device, followed by nmcli connection modify 'Wired connection 2' ipv4.never-default true.
- Solution seems to work, otherwise I couldn't edit this line.
- Should not forget to turn off again when no longer needed.
- Strange, when disconnecting the ethernet cable and reconnecting, I get lights on my LAN, but no wired configuration anymore in my Network configuration.
- Could try to add a keyfile, as suggested in this post.
March 4, 2024
- WSL Ubuntu 22.04 was frozen again, so looked at some tricks from this discussion. The combination of stopping the wsl-service, restarting the LxssManager service and doing wsl --shutdown solved this issue without a reboot (my solution of last week).
January 12, 2024
- Received a request for an Ubuntu 22.04 machine. From my last check, four machines are an option.
- ws9 looks the best option, checking who is using it now. Thomas has used it still last Monday.
Previous Labbooks