@sdetweil
Thank you. I’ve been using ‘npm run start’ as per the documentation off the manual installation instructions.
Just tried ‘npm start’ as well, which also seems to work identically (again, new to npm, so not sure of the difference)

manually exporting the GPU disable at the terminal command line has proven to fix the error for that single run, but after a reboot and not manually entering the export of the GPU disable from command line again, the problem returns.

Seems my solution is starting to boil down to understanding the exact process of how npm launches when issuing the ‘start’ command, and finding the config script?

Fresh install on virgin image of latest Raspberry Pi "Bullseye’ desktop.
Ran the ‘Manual Install’ method:

However, I keep getting the following error, which some of the forum entries here are saying that GPU needs to be disabled in the startup script.

[27.06.2023 05:52.11.136] [LOG]   Launching application.
MESA-LOADER: failed to retrieve device information
MESA-LOADER: failed to retrieve device information
MESA-LOADER: failed to retrieve device information
../../buildtools/third_party/libc++/trunk/include/vector:1425: assertion __n < size() failed: vector[] index out of bounds[8967:0627/055214.116125:ERROR:gpu_process_host.cc(991)] GPU process exited unexpectedly: exit_code=6
MESA-LOADER: failed to retrieve device information
MESA-LOADER: failed to retrieve device information
MESA-LOADER: failed to retrieve device information
../../buildtools/third_party/libc++/trunk/include/vector:1425: assertion __n < size() failed: vector[] index out of bounds[8967:0627/055215.773072:ERROR:gpu_process_host.cc(991)] GPU process exited unexpectedly: exit_code=6
MESA-LOADER: failed to retrieve device information
MESA-LOADER: failed to retrieve device information
MESA-LOADER: failed to retrieve device information
../../buildtools/third_party/libc++/trunk/include/vector:1425: assertion __n < size() failed: vector[] index out of bounds[8967:0627/055217.407482:ERROR:gpu_process_host.cc(991)] GPU process exited unexpectedly: exit_code=6
[9010:0627/055219.369280:ERROR:command_buffer_proxy_impl.cc(128)] ContextResult::kTransientFailure: Failed to send GpuControl.CreateCommandBuffer.
[27.06.2023 05:52.20.355] [LOG]   Create new calendarfetcher for url: http://www.calendarlabs.com/ical-calendar/ics/76/US_Holidays.ics - Interval: 300000

I’m new to npm, so I’m not sure where to look for the ability to disable GPU if that’s the issue.

@sdetweil Thank you, I was all excited there for a second, but it didn’t fix the issue.
As a secondary test, I commented out the actual start line in mm.sh, and the ‘npm run start’ still worked anyways. Looks like the ‘manual installation’ does not use your script.
I might start a new thread here with a more specific question regarding disabling GPU and the location of the start scripts outside of config.js, when using this Manual Installation instruction off the main page.

This is still the result of the run:

@sdetweil Sort of the same problem here. Absolutely fresh install on a new Pi using the manual installation 2 days ago (June 24).
Getting

MESA-LOADER: failed to retrieve device information
../../buildtools/third_party/libc++/trunk/include/vector:1425: assertion __n < size() failed: vector[] index out of bounds[6958:0626/084246.096948:ERROR:gpu_process_host.cc(991)] GPU process exited unexpectedly: exit_code=6

I’m relatively new to npm, and I’m having difficulty locating that script you are referencing (“the script used to start mm”) Where does that reside exactly?
Thanks