Whats your average CPU temp when running on a Pi3?

Tags: #<Tag:0x00007f6170967c50> #<Tag:0x00007f6170967b10> #<Tag:0x00007f61709679d0>

Hey guys,

usually the CPU temperature of my cased Pi 3 (with cooling kit installed) was around 60°C. Recently it went up to >80°C - I removed both the case as well as the WiFi and Z-Wave Stick, but it still goes up to 75°C, which I don’t think is healthy.
My understanding was that the average temp on the Pi 3 should be around 60° - is that the case?
Also did you come across cases that led to such a high CPU temp?


There is no typical average value for this.
What does htop say, are there processes to consume a lot of CPU ?
That’s probably the only reasonable explanation, short of an electrical breakdown of some sort.

I’d say around 60°C is what I’m seeing normally.
I have the standard case and a Razberry on the header.
Check your load average in htop!
Also your ambient temp might be of interest :slight_smile:

My Raspberry 4b 4GB is pretty cool between 38 °C and 40 °C

2020-06-01 07:57:00.217 [vent.ItemStateChangedEvent] - SensorCPUTemp changed from 38.9 °C to 39.4 °C
2020-06-01 07:58:00.219 [vent.ItemStateChangedEvent] - SensorCPUTemp changed from 39.4 °C to 38.9 °C

Housing is a heatsink with double fans.

a020d3 Q1 2018 3 Model B+ 1.3 1 GB (Mfg by Sony)
“Raspbian GNU/Linux 9 (stretch)”
openhab2 2.5.4-1

running OHv2, mosquitto, Stretch lite image
in original rPi case with original power supply

top - 12:49:55 up 31 days, 16:26,  1 user,  load average: 0.30, 0.38, 0.42
Tasks: 125 total,   1 running,  77 sleeping,   0 stopped,   0 zombie
%Cpu(s):  3.2 us,  2.0 sy,  0.0 ni, 94.5 id,  0.0 wa,  0.0 hi,  0.3 si,  0.0 st
KiB Mem :   999036 total,   129480 free,   467496 used,   402060 buff/cache
KiB Swap:   102396 total,    91132 free,    11264 used.   413676 avail Mem

Thanks for your feedback guys.
I checked the past CPU temp in Grafana and noticed that the temperature went up when I activated the amazonechocontrol binding.
I uninstalled the binding and now the temperature is back at ~60°C…

I have several raspis running at home, 2 of them on openhabian, and had heat and power-issues at first that were not noticed. I wrote a little python script to monitor them so I can see their state and health on a habpanel page. You can find it on https://github.com/planetar/pi-health.service

I put a screenshot there showing the temperatures of my raspis floating between 32 and 50 °C with the mean around 40°C

This script is intended to run as a daemon on a raspberry pi with raspbian or a derivative like octopi, openhabian etc. It collects cpu health data including temperature, frequency, throttling, under_voltage, cpu-load, memory use and post them to an MQTT broker. Namely when a number of raspberries are active like in a smarthome setup or with several 3d printers on octoprint this helps to survey the well being of those raspis and integrate their state on a dedicated page, i.e. with openhab2 and the mqtt binding.

I suggest installing rCPU, it monitors all 4 cores and the temperature and displays it via a web api, just point at RaspberryPiIP:8111 and all the info is right there.

It’s been a while since the last update, but as it is it’s solid and reliable. Helps to keep an eye on any problems as they occur.

Thanks for the hint.
So the last week I ran OH without using a case and the temperature was between 60° and 65°.
I just put it back in the case and it went up again to >80°.

I mean it makes sense that it’ll go up, but more than 15° seems a bit weird, doesn’t it…?

Checking htop I don’t see anything suspicious… npm, influxdb, mosquitto, openhab itself… but nothing hogging huge amount of CPU usage.

If you consider the limited space in the case compared to the room around without the case (in %) you migth reconsider if that is weird.
I’m using a case with relatively large openings (openHAB logo!), the installation is inside a wardrobe.
On the logs of the CPU I can tell wether the wardrobe is open and/or if a window in a connecting room is open!

Just did a htop when the temperature was again around 80°, and I saw that mem is almost full and swp is maxed out.

  676 openhab    20 S 100. 43.1  8h05:09 /usr/bin/java -Dopenhab.home=/usr/share/openhab2 -Dopenhab.conf=/etc/openhab2 -Dopenhab.runtime=/usr/share/openhab2/runtime -Dopenhab.userdata=/var/lib/openhab2 -Dopenhab.logdir=/var/log/openhab2 -Dfelix.cm.dir=/var/lib/openhab2/config -Djava.library.path=/var/lib/openhab2/tmp/lib -Djetty.host= -Djetty.http.compliance=RFC2616 -Dorg.ops4j.pax.web.listening.addresses= -Dorg.osgi.service.http.port=8085 -Dorg.osgi.service.http.port.secure=8443 -Djava.awt.headless=true -Dgnu.io.rxtx.SerialPorts=/dev/ttyACM0 -Djava.endorsed.dirs=/opt/jdk/zulu8.40.0.178-ca-jdk1.8.0_222-linux_aarc...
  809 openhab    20 R 100. 43.1  7h40:32 /usr/bin/java -Dopenhab.home=/usr/share/openhab2 -Dopenhab.conf=/etc/openhab2 -Dopenhab.runtime=/usr/share/openhab2/runtime -Dopenhab.userdata=/var/lib/openhab2 -Dopenhab.logdir=/var/log/openhab2 -Dfelix.cm.dir=/var/lib/openhab2/config -Djava.library.path=/var/lib/openhab2/tmp/lib -Djetty.host= -Djetty.http.compliance=RFC2616 -Dorg.ops4j.pax.web.listening.addresses= -Dorg.osgi.service.http.port=8085 -Dorg.osgi.service.http.port.secure=8443 -Djava.awt.headless=true -Dgnu.io.rxtx.SerialPorts=/dev/ttyACM0 -Djava.endorsed.dirs=/opt/jdk/zulu8.40.0.178-ca-jdk1.8.0_222-linux_aarc...

Is there a way to find out what can be causing this?

No standardized.
Show the amount of mem (you unfortunately seem to just have removed that info). It was 1400M or the like. That’s twice as much as it should be (unless you installed 64bit Java, but you didn’t, did you ?), so you likely have a memleak. Slowly remove bindings one by one until mem usage stabilizes.
In general, give way more information upfront if you want help:
How to ask a good question / Help Us Help You - Tutorials & Examples - openHAB Community

Also, your CPU is at 100%. This can be due to swapping but it often also is due to ever-ongoing rule recompilation. Search the forum for that. You will find answers such as that you must not edit items and rules in place.

Yes, it’s 1400M and I am not using 64bit Java.
I’ll try removing bindings… somehow I still think amazonechocontrol is responsible for it… I’ll check back once I know more.

Edit: Thanks for the rule recompilation hint, I’ll look into that as well.

The red one is an octoprint server with a huge (8010) regulated fan but when it’s busy it shows.
OH2 runs on the turquoisie colored, a small fan @ 3V in a very open case. It seems to run cooler after I migrated my rules from DSL to jython. It is a raspi 3B+ while all the others are 3B