@Alvinger Thanks for the quick fix! :)
My first test…
Looks good so far. Too early to tell anything about memory leaks, but I will keep my mirror running and check tonight. Currently electron shows 18% mem used (for the 2 electron processes).
@cyberdie said in Mirror turning black after a while:
@alvinger In my case, the problem started after last MM update.
Same here. Running the latest ResRobot.
But I think others have reported the same problem but with a different module.
@Alvinger Do you know what could be the problem with ResRobot here?
Maybe this can be the cause?
@cyberdie @schlachtkreuzer6 It seems to be related to a certain module. Most likely it is a memory leak you have too. Try to find out by disabling modules until it works.
It would be good to find the root cause to this though as many have problems. But that is a bit out of my expertise.
My problem was MMM-ResRobot. I deactivated it and kept all other modules and now CPU is at 2% and memory stable over time at around 10%. Not sure if the MM/electron/ResRobot update caused the problem, but it worked ok before. Strange…
@Mykle1 @cruunnerr Thanks. Nmon is a great tip. I can clearly see that I have a CPU issue beside the memory leak. CPU for electron process is at 100% all the time. Currently at 30% mem, but I guess this will increase over time. Will check later.
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1146 pi 20 0 700648 300408 47112 R 100.0 31.6 215:22.52 electron
I’ve read the CPU usage thread and can conclude that I use electron v1.4.15 and MM v2.2.0. Mykle1, the thread said to use an older MM version, but I guess the new one should work ok too? Since electron was forced to the older version in the new release?
Now all I have to do is check each module I run and check CPU and mem for each :)
@bimalnair Have you tried to find out which module is causing problems for you yet?
I have finally got a dev log file from the error… As it’s not timestamped, it’s hard to draw any conclusions. But this is one of the lasts logs, which presumably is when the mirror goes black:
[5966:0125/191204:FATAL:memory_linux.cc(35)] Out of memory.
#0 0x0000012a3bb0
#1 0x0000012b4406
#2 0x0000012c8ae6
#3 0x000000d8bb1c
#4 0x000000d8ba9e operator new()
#5 0x0000012b9ee0
#6 0x0000012b94ce
#7 0x0000012ba204
#8 0x0000012cb2c6
#9 0x0000012b89f6
#10 0x000000b58d18
#11 0x000000a34c38
#12 0x000000a35a80
#13 0x000000a349c2
#14 0x00000221e2fa main
#15 0x000074c0f294 __libc_start_main
I did not check which process is eating all the memory though. That will be the next step.
Other than that I can see in the log that I lost internet connection a couple of times:
fatal: unable to access 'https://github.com/MichMich/MagicMirror.git/': Could not resolve host: github.com
Whoops! There was an uncaught exception...
TypeError: Cannot read property 'feed' of undefined
at ClientRequest. (/home/pi/MagicMirror/modules/MMM-Skolmaten/node_helper.js:45:20)
And a lot of these logs, which I guess is ok:
Reusing 14 cached departure(s) for module: MMM-ResRobot
Any bright ideas? :)
@Henrik From what I’ve read, a collaborator has write access directly to the repo. You can push code directly as if you are the owner.
@erq @Henrik Thanks for the tip! I have made a quick solution for now… changed package.json to have dependency to rss-parser v2.12.0.
Checked the rss-parser readme and the latest version (3) has changed it’s interface (“A few minor breaking changes were made in v3…”). If I find the time, I will update my code to support the latest version.
