Read the statement by Michael Teeuw here.
Mirror Freezing
-
I have updated everything trying to get past this.
I installed a real time clock.
Fresh install of Debian 12
Fresh install of MagicMirror and restore with Sam’s backup/restore script.
So everything is fresh. I am running this on a NVME ssd stick.
very fast.
Ok since i gave you all the background.
Here is the problem I am having.
MM will run fine for hours, then all the sudden MM freezes, the time stops and it stays on the page that the time freezes on.
I have about 40 different modules running and they run great until the freeze up happens.
Below is the errors I am getting in the pm2 logs.0|mm | [2048:0303/182231.302189:ERROR:gles2_cmd_decoder_passthrough.cc(928)] ContextResult::kFatalFailure: fail_if_major_perf_caveat + swiftshader 0|mm | [2048:0303/182231.339696:ERROR:gles2_cmd_decoder_passthrough.cc(928)] ContextResult::kFatalFailure: fail_if_major_perf_caveat + swiftshader 0|mm | [2025-03-03 18:22:33.874] [ERROR] We expected a redirect to guce.yahoo.com, but got https://finance.yahoo.com/quote/AAPL/ 0|mm | [2025-03-03 18:22:33.875] [ERROR] We'll try to continue anyway - you can safely ignore this if the request succeeds 0|mm | /home/pi/MagicMirror/node modules/electron/dist/electron exited with signal SIGINT 0|mm | [3690:0304/064512.174017:ERROR:gbm_wrapper.cc(74)] Failed to get fd for plane.: No such file or directory (2) 0|mm | [3690:0304/064512.174366:ERROR:gbm_wrapper.cc(257)] Failed to export buffer to dma_buf: No such file or directory (2)These repeat over and over for 20 or so times. I didnt count them.
-
@plainbroke i think most often is the parent dies but the child was still running…
-
Oh and I can ssh in and run date and the time is correct.
If I run MM in a separate browser on my laptop I pops up and runs fine and very clean. Time is correct. ect. which then causes the MM on the wall to start working again…
I did run your update script with force tonight when I came home and it had been frozen for approx. 3.5 hrs.
I am currently watching to see if it freezes again. so frustrating. -
@plainbroke start in server mode and connect chrome or some other browser, sounds like an electron problem
maybe change the pages rate a little (10%)
-
@also to stackoverflow
which points to a graphical module, globe, something doing heavy charts.
-
@sdetweil
you are correct electron is hoarding resources.
Is there a way to kill a PID in debian to see what module may be causing this.
1998
1988
and 1949 are top three cpu hogs… -
@plainbroke you can
sudo kill -9 pid
but i dont know what functions are in what process, electron decides that. one is the page/tab
-
@sdetweil
Just did a kill 1998 and it is now not showing up but the mirror is still working,
I guess I will wait until one of the modules doesn’t work anymore.
Then i will know which one to remove, right?
Turns out that was the wireless connection I think.
MM stopped and I can not access it through ssh time to POPO.
I will keep after this but I will also stop globe in the config.js. just to see if that helps.
Thanks for your help I have learned alot tonight about the way debian and MM run. -
@plainbroke i think page switching while globe is drawing
is probably the cause -
@plainbroke this is the best Answer
-
@sdetweil
Not so sure that it was globe came home today and the same screen was stopped,
MMM-Dogs and the time was the same 3:37 pm but the real time was 4:30 pm.
So re-enabled globe and disabled MMM-Dogs to see if it does it again tomorrow…
I am almost wondering if it isn’t something to do with MMM-birthday-paged.
It was the last module I remember adding since this started.
I will slowly figure it out by disabling one module at a time that seems to hog resources.By the way what do it mean by zombie in top.?
top - 16:53:44 up 17 min, 3 users, load average: 5.34, 5.15, 3.71 Tasks: 222 total, 3 running, 218 sleeping, 0 stopped, 1 zombie %Cpu(s): 77.3 us, 12.1 sy, 0.0 ni, 10.6 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st MiB Mem : 7820.9 total, 4021.5 free, 1923.3 used, 2830.4 buff/cache MiB Swap: 512.0 total, 512.0 free, 0.0 used. 5897.6 avail MemSlow learner. But trying anyways.
MM is on Raspberry Pi 4B w/8gb ram loaded on a 128gb nvme drive.
Running Trixie and the latest MM version. -
-
-
@plainbroke i think most often is the parent dies but the child was still running…
-
@sdetweil @karsten13
Found Itps ux | awk '{if($8=="Z") print}' pi 1170 0.0 0.0 0 0 ? Z 16:36 0:00 [xsettingsd] <defunct>What is xsettingsd?
Should x11 be running with wayland ?
Seems like a duplicated background app running or do they not do basically the same thing.?I did kill the 1166 which turned out to be the parent of the 1170 xsettingsd
Mirror restarted I assume due to PM2 being involved.
It restarted MUCH faster than normal. The MMM-Snow is running smoother.
I guess only time will tell. -
P plainbroke has marked this topic as solved on
-
P plainbroke has marked this topic as solved on
Hello! It looks like you're interested in this conversation, but you don't have an account yet.
Getting fed up of having to scroll through the same posts each visit? When you register for an account, you'll always come back to exactly where you were before, and choose to be notified of new replies (either via email, or push notification). You'll also be able to save bookmarks and upvote posts to show your appreciation to other community members.
With your input, this post could be even better 💗
Register Login