Script temperature logger | How to figgure out the system temperatures

Hello all,

Currently I'm running Unraid V6.12.14 with the apps:

- ITE IT87 Driver

- Nuvoton NCT6687 Driver

- Dynamix System Autofan

- Dynamix System Temp

I have searched for directories related to hardware info, e.g. the temperatures which are displayed in the dashboard:
 

```bash
find /sys -type d -name "*hw*"
find: ‘/sys/kernel/slab’: Input/output error
/sys/class/hwmon
/sys/devices/platform/it87.2592/hwmon
/sys/devices/platform/it87.2592/hwmon/hwmon3
/sys/devices/platform/coretemp.0/hwmon
/sys/devices/platform/coretemp.0/hwmon/hwmon2
/sys/devices/pci0000:00/0000:00:1d.1/0000:06:00.0/nvme/nvme0/hwmon1
/sys/devices/virtual/thermal/thermal_zone0/hwmon0
/sys/devices/virtual/misc/hw_random
/sys/module/hwmon_vid
```

I am wondering where the temperatures of the disks in the array are stored. Anyone has an idea? I don't want to read out the specs using smartctl, since the values should be already available somewhere.

Why I am doing this?

I want to record the temperatures of system, in order to optimize the airflow/cooling within the case. Therefore I have wrote a little script, which selects the specific temperatures of my components:

#!/usr/bin/bash
# Version 2
# Better performing temperature logger than logging
# with $(sensors | awk '/CPU Temp/ {print $3}').
# This script gets the current temperature values
# and sleeps (sleep_interval) for 10s
# Usage:
# bash temp_loggerV2.sh <test_period> <specs>
#
# Example, logging for 60 seconds and specifying
# an idle system:
# bash temp_loggerV2.sh 60 idle
#
# Tested on Unraid V6.12.14

# Operants
test_period=$1
specs=$2
sleep_interval=10
if ((test_period<sleep_interval));then
	test_period=$((10/sleep_interval))
else
	test_period=$((test_period/sleep_interval))	
fi
logdir=/boot/config/plugins/stresstest

# Variables
path_T_cpu="/sys/devices/platform/it87.2592/hwmon/hwmon3/temp2_input"
path_T_main="/sys/devices/virtual/thermal/thermal_zone0/hwmon0/temp1_input"
path_T_nvme="/sys/devices/pci0000:00/0000:00:1d.1/0000:06:00.0/nvme/nvme0/hwmon1/temp1_input"
T_cpu=0
T_main=0
T_nvme=0
sum_T_cpu=0
sum_T_main=0
sum_T_nvme=0

# Messages
msg_head="Measurement\tT_cpu / (k°C)\tT_main / (k°C)\tT_nvme / (k°C)\n"
msg_line="==============================================================\n"

# Verifying stresstest log directory & creating logfile
if [[ ! -d "$logdir" ]]; then
  mkdir "$logdir"
fi
log_filename=${logdir}/$(date +%Y%m%d_%H%M%S)_temperature_${specs}.log
touch $log_filename

# Temperature logger function
record_temps() {
	test_period=$1
	printf "${msg_head}" | tee -a ${log_filename}
	printf "${msg_line}" | tee -a ${log_filename}
	for ((i = 0 ; i < "${test_period}" ; i++)); do
		# CPU
		T_cpu=$(cat $path_T_cpu)
		sum_T_cpu=$(echo ${sum_T_cpu} ${T_cpu} | awk '{print $1 + $2}')

		# Mainboard
		T_main=$(cat $path_T_main)
		sum_T_main=$(echo ${sum_T_main} ${T_main} | awk '{print $1 + $2}')

		# NVME
		T_nvme=$(cat $path_T_nvme)
		sum_T_nvme=$(echo ${sum_T_nvme} ${T_nvme} | awk '{print $1 + $2}')

		# Logging
		printf "${i}\t\t${T_cpu}\t\t${T_main}\t\t${T_nvme}\n" | tee -a ${log_filename}
		sleep $sleep_interval
	done
	# Average
	avg_T_cpu=$(awk -v var1=$sum_T_cpu -v var2=$test_period 'BEGIN { print  ( var1 / var2 / 1000 ) }')
	avg_T_main=$(awk -v var1=$sum_T_main -v var2=$test_period 'BEGIN { print  ( var1 / var2 / 1000 ) }')
	avg_T_nvme=$(awk -v var1=$sum_T_nvme -v var2=$test_period 'BEGIN { print  ( var1 / var2 /1000 ) }')
	printf "${msg_line}" | tee -a ${log_filename}
	echo "Average CPU temperature: ${avg_T_cpu}°C" | tee -a ${log_filename}
	echo "Average Mainboard temperature: ${avg_T_main}°C" | tee -a ${log_filename}
	echo "Average NVME temperature: ${avg_T_nvme}°C" | tee -a ${log_filename}
}

record_temps ${test_period}
printf "Log was written to $log_filename\n"

Anyone who want to use it, has to change the variables path_T_cpu, path_T_main and path_T_nvme according to his/her system. Running it via User scripts destroys the tab seperators, but the log is still readable. The log is stored under /boot/config/plugins/stresstest/<YYYYMMDD_HHMMSS>_temperature_<specs>.log.