MM 2.18 not starting, crashes on startup [exit_code=512]

p1lspeda

Hi,
I ran into a really annoying problem.
All I wanted to do, was update a newsfeed. Unfortunately it went all downhill from there.
What I have done so far:

fresh install
manually installed missing modules via npm e.g. npm install request
adjusted permissions for chrome sandbox

This all has not resulted in success.
As my MM wouldn’t boot up into “normal” mode, i.e. MM shows up with all the bells and whistles (via pm2), I did the following:

I disabled all modules and started adding one module at a time, starting with just the clock
I started MM in servermode
connected via VNC to my MM
opened chrome and opened 127.0.0.1:8080
This resulted in a MM interface with a clock.
My MM seems to generally be working.

I went through and added one module at a time and repeated above mentioned steps.
Last module, that caused trouble was MMM-CalendarWeek.
This could be a coincidence though.

The pm2 log states

pi@magicmirror:~ $ ./MagicMirror/run-start.sh
[16.02.2022 20:04.41.387] [LOG]   Starting MagicMirror: v2.18.0
[16.02.2022 20:04.41.405] [LOG]   Loading config ...
[16.02.2022 20:04.41.418] [LOG]   Loading module helpers ...
[16.02.2022 20:04.41.422] [LOG]   No helper found for module: MMM-Cursor.
[16.02.2022 20:04.41.425] [LOG]   No helper found for module: clock.
[16.02.2022 20:04.41.866] [LOG]   Initializing new module helper ...
[16.02.2022 20:04.41.868] [LOG]   Module helper loaded: MMM-PublicTransportHafas
[16.02.2022 20:04.41.993] [LOG]   Initializing new module helper ...
[16.02.2022 20:04.41.996] [LOG]   Module helper loaded: MMM-Fuel
[16.02.2022 20:04.42.057] [LOG]   Initializing new module helper ...
[16.02.2022 20:04.42.059] [LOG]   Module helper loaded: MMM-Weather
[16.02.2022 20:04.42.060] [LOG]   No helper found for module: MMM-Year-Progress.
[16.02.2022 20:04.42.584] [LOG]   Initializing new module helper ...
[16.02.2022 20:04.42.586] [LOG]   Module helper loaded: MMM-Holiday
[16.02.2022 20:04.43.072] [LOG]   Initializing new module helper ...
[16.02.2022 20:04.43.074] [LOG]   Module helper loaded: MMM-DWD-Pollen
[16.02.2022 20:04.43.117] [LOG]   Initializing new module helper ...
[16.02.2022 20:04.43.119] [LOG]   Module helper loaded: newsfeed
[16.02.2022 20:04.43.120] [LOG]   All module helpers loaded.
[16.02.2022 20:04.43.292] [LOG]   Starting server on port 8080 ...
[16.02.2022 20:04.43.317] [LOG]   Server started ...
[16.02.2022 20:04.43.319] [LOG]   Connecting socket for: MMM-PublicTransportHafas
[16.02.2022 20:04.43.321] [LOG]   Connecting socket for: MMM-Fuel
[16.02.2022 20:04.43.323] [LOG]   Starting module helper: MMM-Fuel
[16.02.2022 20:04.43.324] [LOG]   Connecting socket for: MMM-Weather
[16.02.2022 20:04.43.326] [LOG]   Connecting socket for: MMM-Holiday
[16.02.2022 20:04.43.327] [LOG]   Starting node_helper for: MMM-Holiday
[16.02.2022 20:04.43.329] [LOG]   Connecting socket for: MMM-DWD-Pollen
[16.02.2022 20:04.43.330] [LOG]   MMM-DWD-Pollen helper started ...
[16.02.2022 20:04.43.331] [LOG]   Connecting socket for: newsfeed
[16.02.2022 20:04.43.332] [LOG]   Starting node helper for: newsfeed
[16.02.2022 20:04.43.333] [LOG]   Sockets connected & modules started ...
[16.02.2022 20:05.07.955] [LOG]   Launching application.
**[2139:0216/200529.312115:ERROR:gpu_process_host.cc(968)] GPU process exited unexpectedly: exit_code=512
[2139:0216/200529.355982:ERROR:network_service_instance_impl.cc(916)] Network service crashed, restarting service.
[2283:0216/200530.410224:ERROR:viz_main_impl.cc(161)] Exiting GPU process due to errors during initialization
[2309:0216/200530.708608:ERROR:sandbox_linux.cc(376)] InitializeSandbox() called with multiple threads in process gpu-process.
[2259:0216/200530.832997:ERROR:command_buffer_proxy_impl.cc(125)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.```**

######################

pi@magicmirror:~ $ npm -v
8.3.1
pi@magicmirror:~ $ node -v
v16.14.0

######################

pi@magicmirror:~ $ cat /etc/os-release
PRETTY_NAME="Raspbian GNU/Linux 11 (bullseye)"
NAME="Raspbian GNU/Linux"
VERSION_ID="11"
VERSION="11 (bullseye)"
VERSION_CODENAME=bullseye
ID=raspbian
ID_LIKE=debian
HOME_URL="http://www.raspbian.org/"
SUPPORT_URL="http://www.raspbian.org/RaspbianForums"
BUG_REPORT_URL="http://www.raspbian.org/RaspbianBugs"

I really have no idea what else I can do.
Any help is much appreciated

Thanks
P

sdetweil

@p1lspeda see
https://forum.magicmirror.builders/topic/15778/fix-for-black-screen-in-2-16-and-later

p1lspeda

@sdetweil
Thanks for the link (I had already searched through that topic).
I can’t find anything related there though to my problem.
I cannot find any clues in regards to my ERROR

[2139:0216/200529.312115:ERROR:gpu_process_host.cc(968)] GPU process exited unexpectedly: exit_code=512
[2139:0216/200529.355982:ERROR:network_service_instance_impl.cc(916)] Network service crashed, restarting service.
[2283:0216/200530.410224:ERROR:viz_main_impl.cc(161)] Exiting GPU process due to errors during initialization
[2309:0216/200530.708608:ERROR:sandbox_linux.cc(376)] InitializeSandbox() called with multiple threads in process gpu-process.
[2259:0216/200530.832997:ERROR:command_buffer_proxy_impl.cc(125)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.```**

All my modules load correctly

p1lspeda

@sdetweil
Ran my config (I just commented the various modules within the config) with just the default “clock” module and still got the Error msg,
Does that prevent them (= the unwanted modules) from being loaded ?

pi@magicmirror:~/MagicMirror $ ./run-start.sh
[17.02.2022 06:27.16.057] [LOG]   Starting MagicMirror: v2.18.0
[17.02.2022 06:27.16.075] [LOG]   Loading config ...
[17.02.2022 06:27.16.085] [LOG]   Loading module helpers ...
[17.02.2022 06:27.16.090] [LOG]   No helper found for module: MMM-Cursor.
[17.02.2022 06:27.16.094] [LOG]   No helper found for module: clock.
[17.02.2022 06:27.16.096] [LOG]   All module helpers loaded.
[17.02.2022 06:27.16.292] [LOG]   Starting server on port 8080 ...
[17.02.2022 06:27.16.318] [LOG]   Server started ...
[17.02.2022 06:27.16.320] [LOG]   Sockets connected & modules started ...
[17.02.2022 06:27.16.736] [LOG]   Launching application.
[17158:0217/062738.598660:ERROR:gpu_process_host.cc(968)] GPU process exited unexpectedly: exit_code=512
[17158:0217/062738.647202:ERROR:network_service_instance_impl.cc(916)] Network service crashed, restarting service.
[17230:0217/062739.406765:ERROR:viz_main_impl.cc(161)] Exiting GPU process due to errors during initialization
[17256:0217/062739.710018:ERROR:sandbox_linux.cc(376)] InitializeSandbox() called with multiple threads in process gpu-process.
[17200:0217/062739.830940:ERROR:command_buffer_proxy_impl.cc(125)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.

p1lspeda

@sdetweil
I’ve done some more investigating.
After running the run-start.sh script and noticing that MM doesn’t start as intended, I stop MM by pressing ctrl+c within the terminal.
I can see via VNC that “Electron” shuts down.
So I assume that my problem has something to do with electron

pi@magicmirror:~ $ cd MagicMirror
pi@magicmirror:~/MagicMirror $ ./node_modules/.bin/electron --version
v16.0.9

See screenshot:

Hope this helps trying to figure out what I going on.
Thanks
P

p1lspeda

@sdetweil
This is really weird, trying to launch chromium (as in your run-start.sh) give’s me an Error

pi@magicmirror:~ $ chromium-browser
[0217/142409.967126:ERROR:elf_dynamic_array_reader.h(61)] tag not found

sdetweil

@p1lspeda see this post, for a possible fix

https://forum.magicmirror.builders/topic/10322/flashing-and-green-lines-after-couple-of-minutes/2?_=1645106186637

p1lspeda

@sdetweil
Hi,
things are getting better (but still really weird).
I disabled the OpenGL driver and now my MM is 96% back up and running.
MMM-CalendarWeek causes MM not to show up.
After disabling it within the config.js, its up and running.

Thanks for all your help!

sdetweil

@p1lspeda see that link I posted above about blank screen