default calendar fetching issues

[nowayto]

@sdetweil
behaviour seems the same. I suspect this, could this issues be related to the raspberry pi 3 computing capacity? could it be not enough?

Is there a way to delay the pm2 automatic startup.
For example I though to try a delay of 1-2 minutes so:
1- raspbian startup
2- delay of 1-2 minutes to make sure all the services are up and the cpu is idle
3- mm startup
4- if it’s possible, I will add a delay to the modules startup also, maybe 20 seconds between every module loading

Honestly I checked his capacity a lot of times using the htop, free and cpu temperature command.
With my amazement I see everytime an amount of free memory, low cpu load and good cpu temp (like 40 to 50 degrees)

But if there is a way to delay everything, I’ll just check if with a “calmly” startup It will go better

[sdetweil]

@nowayto i do not know on pm2

you could put a sleep 120
in the mm.sh to wait a while

how is swap space?

free -m

pi02w only has 512mb

[nowayto]

@sdetweil
Yes I agree, the memory is very limiting.
Although mine is a Pi3 model A, but the memory is 512mb like the pi02w.
When I started the project I had something different in my mind and now I’m pushing it to the limit.

But as I said I checked everytime the CPU temperature, memory and generally resources usage from “htop” command.

This morning MM started with calendar error, I did a “pm2 restart mm” and goes OK.

Now I put “sleep 120” inside mm.sh as first line
I restarted it now

Memory BEFORE restart was this
total used free shared buff/cache available
Mem: 427 156 90 70 179 147
Swap: 1023 231 792

Not so good but not so bad, same time.

I’ll check if this delay could help.
Is there a way to put a little delay between the loading of every module ? Maybe 5/10 seconds, I though it could be very helpful on a limited machine like mine :)

[sdetweil]

@nowayto there is no way to delay loading

[emlowe]

I had considerable problems with running on a model A - but I found dietPI (https://dietpi.com/) to be a solid distro that works well even when using the default electron GUI front-end. (Previously OOM’d consistently)

Note, also I frequently have trouble with the calendar events appears as there seems to be some problem in the websocket communication layer (this bug: https://github.com/MagicMirrorOrg/MagicMirror/issues/3380)

I basically use MMM-RemoteControl to force a refresh in a daily cron job (curl http://localhost:8080/api/refresh) and this works. Why? No One knows

[nowayto]

@emlowe I have a similar problem, I solved modifing mm.sh with

sleep 30
DISPLAY=:0 npm start
sleep 60
pm2 restart mm

“pm2 restart mm” does a refresh like you do with RemoteControl

[sdetweil]

@nowayto if pm2 is being used, its JOB is to keep the app running

so the pm2 restart seems odd as if the app was ended , ctrl-q
pm2 would do restart already