Roswell System Narrative

2023 October 4

Despite the adventures yesterday, no problems overnight.  We'll see.

Updated the Fourmilab_Mirror backup of all Roswell file systems to take
back to Fourmilab as a reference of the state of the machine.
3 Likes

2023 December 15

Applied all updates to packages accumulated in my absence.  Updated 189
packages, including the kernel, Nvidia drivers, and just about
everything else.

Ran a "snap refresh" to update all snaps, of which 11 were updated.

Ran an "apt autoremove" to clean up old kernel and related packages.

Rebooted.  We are now running on kernel:
    6.2.0-39-generic #40~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC.

Updated the "kept back" packages using the procedure from Ragnar2:
   1. apt list --upgradable
   2. apt-get install 
   3. apt list --upgradable
   4. If packages remain, go to 2.
   5. apt-get autoclean
   6. apt-get update
   7. apt-get dist-upgrade
This updated:
    python3-update-manager
    systemd-hwe-hwdb

Rebooted.  An "apt-get upgrade" now reports nothing to do.

Switched to the Windows side.  Applied all updates queued by Windows
Update.  This took about an hour and a half with around six reboots
required in the process.  At the end, there was a scary sequence of
"Updating Firmware" messages  but, after all were done, the thing
came up and appears to work.  Then it presented me with a series of
obtrusive, insulting, and manipulative pop-ups trying to sell their
subscription Orifice crap, cloud storage, and to try to trick me into
re-enabling their garbage spyware browser Edge as the default browser.
Blue Screen of Apu is clearly in the driver's seat and Microsoft has
added to their traditional characteristic of having no pride the
additional distinction of having no shame.

Updated Firestorm Viewer to current version.

Updated Chrome browser to current version.

Updated Steam to current version.

Booted back into Xubuntu to get real work done.
5 Likes

2023 December 16

After having turned off the auxiliary monitor the night before, when
I turned it back on the next day, it was totally blank and activating
the primary screen did nothing.  The monitor said it was receiving an
HDMI signal, and it showed as active on the Settings/Display panel, but
it appeared stuck in screen blank screen saver.  Changing the monitor
configuration in the Display panel did nothing, including designating
the second monitor as primary and back.  I unplugged the HDMI cable,
which dropped the monitor from the Display panel, and when I plugged
it back in, it came alive. I then reset to be right of the main screen
and everything appeared normal again.

Ookla speedtest.net reports, having selected a test to the Manx Telecom
server in Douglas, 16.13 Mbps down, 1.05 Mbps up, 40 ms ping.  This is
just about identical to the Galaxy Tablet S9 on WiFi, so what we're
seeing here is the DSL speed limitation.

Mozilla Firefox has now gotten into the stupid game with Google Chrome
of suppressing the Xfce window title bar, which prevents a user from
dropping down the menu that controls such functions as moving a window
to another workspace or making it appear on all workspaces (see the
entry for 2023-09-27 for a description of this regarding Chrome).
Unlike Chrome, I have yet to find a way to turn off this idiot mode and
restore the system title bar.  But I did discover that you can grab a
Firefox window by its title bar and drag it from the edges of the two
screen pair onto an adjacent workspace, which provides a way to move
the window among workspaces, albeit more tedious, and still doesn't
provide a way to access the other functions of the workspace menu.

Oh, OK, it turns out you can get to the workspaces menu by putting the
mouse in the window and pressing Alt+Spacebar.  Isn't that intuitive?

There's another odd phenomenon with the mouse.  Occasionally (I haven"t
figured out what is the trigger, if anything), the mouse will cease to
allow movement onto the auxiliary screen and windows dragged toward it
disappear at the edge of the primary screen.  I have been able to fix
this by using the Settings/Display panel to change the screen layout
to one above the other, letting it swap to that mode, then revert to
the original configuration before confirming.  Then it's all better
until the next time.  I suspect this is what happened when I found
the auxiliary screen stuck on black when I started today.  It was the
same thing, but occurred while in screen saver.

Trying to get X-Plane 12 running from Steam, it crashes instantly after
being started with the following in
    ~/snap/steam/common/.local/share/Steam/steamapps/common/X-Plane\ 12\Log.txt
        Vulkan Layers       : VK_LAYER_VALVE_steam_overlay_32 (1), VK_LAYER_VALVE_steam_fossilize_64 (1), VK_LAYER_VALVE_steam_fossilize_32 (1), VK_LAYER_VALVE_steam_overlay_64 (1), VK_LAYER_MANGOHUD_overlay (1), VK_LAYER_MANGOAPP_overlay (1)
        0:00:00.000 W/GFX/VK: Unknown device vendor ID 8086, for device Intel(R) Graphics (RPL-S) (skipping)
        0:00:00.000 E/GFX/VK: Failed to find a suitable device. This is what was found:
        0:00:00.000 E/GFX/VK:   NVIDIA GeForce RTX 4090 Laptop GPU reported: Adapter Unique ID doesn't match OpenGL bridge device
        0:00:00.000 E/GFX/VK:   Intel(R) Graphics (RPL-S) reported: Your graphics card or drivers cannot run X-Plane, because they don't support one or more required features. Please try updating your graphics drivers.
        shaderStorageImageMultisample
        0:00:00.000 E/GFX/VK: Picked bridge device: 4.6 (Compatibility Profile) Mesa 23.2.1 refresh - kisak-mesa PPA
        0:00:00.000 E/GFX: Failed to initialize Vulkan
        0:00:00.000 E/GFX: Vulkan layer reported: Adapter Unique ID doesn't match OpenGL bridge device
        0:00:00.000 E/SYS: MACIBM_alert: X-Plane failed to initialize Vulkan and can't run.
        0:00:00.000 E/SYS: MACIBM_alert: Adapter Unique ID doesn't match OpenGL bridge device

According to various documents, this may be due to missing OpenGL
drivers.  They recommend installing:
    apt-get install libglu1-mesa-dev freeglut3-dev mesa-common-dev

I installed these and dependencies and the error pop-up went away.  It
was replaced an a nearly-instantaneous silent crash which did not even
touch the Log.txt file.
4 Likes

I recently ran into this with the latest version of Thunderbird. It’s controlled by a checkbox in the app’s general settings, under “Window Layout”, with the label “Hide system window titlebar”. Which defaulted to on when it first appeared.

Haven’t hit it on my Firefox install yet, but I gather it’ll be soon.

Anyways, I suspect you can kill off this innovation.

3 Likes

2023 December 17

Further experiments with the maddening "Auxiliary display doesn't
unblank" problem...

The first thing I established is that when you're in that state, the
actual problem is that the auxiliary display is blanked.  It is on,
receiving HDMI signals, and the display manager sees it in the correct
position and will let you drag windows partially or entirely onto it
as if it were working normally.  Moving the mouse onto the aux display
and pressing keys does not wake it up (I had thought maybe it wasn't
coming out of screensaver due to lack to stimulus, but no joy).

Using Settings/Power Manager/Display/System/Plugged in I set:
    Laptop lid:
        When laptop lid closed: Switch off display
and with this mode in effect, closing and re-opening the lid seems to
wake up the aux display.  The first time I tried it, I had to
re-establish the display orientation, but after that it seemed to
remember.  We'll see how it behaves when it craps out the next time.
4 Likes

2023 December 18

Installed:
    apt-get install netpbm

2023 December 19

After running idle overnight, when I went back to the system the fan
was running flat-out but there didn't appear to be unusual heat coming
from the vents nor was the (closed) room warm.   Top showed a few
Chrome processes running at 20% of the CPU (but recall this is
core-wise, so that's some pegged cores), but with a virtual memory size
of 33 Gb.  When I killed the top process, the Chrome tab that died
was just the display of Subscription videos.  I killed Chrome with
killall and it terminated normally, but the fan was unaffected.  Top
then showed 99.7 idle, but the fan continued blasting.  Xorg showed a
virtual sixe of 28.5g with 230504 res.  Overall memory status was:
    MiB Mem :  63977.2 total,  16226.6 free,   3452.2 used,  44298.5 buff/cache
    MiB Swap:   2048.0 total,   2048.0 free,      0.0 used.  57944.1 avail Mem

I logged out and back in, but nothing changed.  The fan kept running
and Xorg continued to show 28.5 Gb.  I then rebooted.  At the start
of the reboot process "Support Assistant" popped up and proceeded to
run tests on system fans.  Right at the end of "Testing CPU fan" (or
whatever--I could not screen shot it) and at the start of "Testing
video fan", the runaway fan stopped.  It proceeded through more fan
tests, pronounced the system healthy, and completed the reboot normally.
Now that the system is back up, Xorg still shows 28.3 Gb virtual size.
This must be some kind of sparse allocation because next to nothing
using the window system is running right now.

I'm still not sure of everything that causes the lock up of the aux
screen, but one thing that seems to cause it every time is the display
going into the screensaver.  When it comes out, the aux screen always
remains black until you close and re-open the lid.
2 Likes

2023 December 20

Paired the built-in Bluetooth interface with the EPOS ADAPT 460T
headset.  I launched the Settings/Bluetooth Devices panel, then put
the headset into pairing mode.  After a couple of power cycles on the
headset and attempts to get it in pairing mode, the headset showed up
in the panel.  I then selected it and clicked the "keyhole" icon to
initiate pairing.  This was a fiddly process, but eventually it paired
the the audio output switched from the headphone jack to Bluetooth.

This worked well, but if I turn on any other device (phone or tablet)
which is also paired to the headset, it ceases to receive audio from
the computer until the other device is turned off and power is cycled
on the headset.

I have not tried using the microphone on the headset.
1 Like