This program appends datetime-stamped recordings of Raspberry Pi CPU core temperatures as rows in a sqlite3 database. The program uses (creates if necessary) the sqlite3 database, creates a table for the data if necessary, and appends rows at regular periodic intervals. The data can be reported via sqlite3 commands or displayed graphically by a web page or html file (generated by programs included in the package).
Use sqlite3 to view the data, e.g.:
$sqlite3 /var/databases/MyPiTemps.db
sqlite> select * from PiCoreTemps where DateTime > datetime('now', '-48 hours');
sqlite> .mode line
sqlite> select count(cputemp),avg(cputemp),max(cputemp),min(cputemp) from PiCoreTemps where datetime>datetime('now', '-24 hours');
Alternatively, to get a summary of the temperature readings for the past 24 hours, use the sqlite3 provided (installed in /usr/local/bin):
$sqlite3 /var/databases/MyPiTemps.db < /usr/local/bin/CheckTemps.sqlite3
The PHP-generated web page is suitable as the foundation for a web site page that displays a graphical presentation of the time/temp data. The Python program generates an html file for graphical presentation, and that file can be viewed with an html browser. The prototype web PHP code and the Python program use the same Google Charts tools and so generate equivalent displays.
The data logger will acquire data, and you can obtain formatted text reports on that data, with just the sqlite3 packages installed (described below).
If you intend to display the results graphically, you'll need to either be running Apache2 (if you want to display the graphs on your own web pages) or you'll generate an html file with the Python program included here and then view the html file with a browser. If you haven't done so, you might install Apache2 in preparation for using this package. The Apache2 installation on Raspbian is well documented and won't be repeated here.
The PHP code that pulls the data from the sqlite3 database and formats it for the Google Charts graphing system requires the php-sqlite3 module. Even if you have Apache2 running and have installed the sqlite3 packages, you might not have installed the interface module. Prepare for graphical display by installing that package: sudo apt-get install php5-sqlite.
The graphing programs use the Google Charts system for doing the actual graphing, and that system in invoked at the time those programs are run, so no package installation is required for those. The Makefile installation is set to install the logging program as a daemon so that it is automatically started at boot time.
- Program compilation/linking requires packages for sqlite3 and libsqlite3-dev. Use
sudo apt-get install sqlite3 libsqlite3-devto install those if you haven't already. - Connect to install directory and type
maketo simply compile and link to create the executable data logger. Or compile withgcc -o RPiTempLogger -lsqlite3 RPiTempLogger.c - Type
sudo make installto move the resulting executable to "/usr/local/bin", set it up to start as a daemon at reboot, and copy the graphing code to "/var/www". If you're running on openSUSE, the systemd directories are different from Raspbian: you'll need tosuto get a root prompt and thenSUSE=yes make install.
Review "Makefile" for more detailed instructions, including options for changing the names and locations of files.
The C program contains several compile parameters that can be changed, too:
- DelaySec: time between samplings, in seconds, as an integer [default 300]
- DBName: path and filename of the database file in which the data will be stored [default is "MyPiTemps.db" in "/var/databases", that is created upon install]
Run as ./RPiTempLogger to test the program. It starts by sampling an printing the temperature and adding the sample data to the sqlite3 database; it repeats that (every 5 minutes, by default) until terminated. Run as ./RPiTempLogger & to run detached, which it will do until a reboot. Do sudo make install to install RPiTempLogger as a service that starts at boot time under systemd (see variation above for openSUSE install). Or include the command /usr/local/bin/RPiTempLogger in a startup script to run as daemon automatically whenever the system is rebooted.
When the program is running, it doesn't keep the database file open: when the program wakes up periodically, it samples the data, opens the database, appends the sample reading, and closes the database. Exposure to table corruption as a result of a Pi restart or crash is minimized.
You can generate a graphical representation of the data using the python program "CPUTemps.py" that is copied during install to "/var/www/cgi-bin/" and use it as a displayable web page on your Pi's web site if you're running Apache. The file index-RTL.php is copied to "/var/www" as a prototype for a web site home page that will display the data graphically.
Written by HDTodd, Williston Vermont, November, 2015.
Updated README.md documentation and Makefile for installation under systemd and to support installation on openSUSE in March, 2018.
hdtodd@gmail.com