A collection of scripts and resources used in a Bachelor's study on Stellar energy consumption in the spring of 2021.
For the final report, see its page on Lund University Publications. There, you can read more about the methodology.
The following scripts were used to monitor a Stellar Validator node, with the
most important part being what is contained in
power-monitoring.
cfg- Config filesplots- Scripts used for creating plots and tables. See subdir for README.power-monitoring- systemd service for monitoring power consumptionsysstat-montioring- systemd service for monitoring system resource usage
The power monitoring is managed as a system service, which calls the script
manager.sh. The script will read the power
consumption of the machine during approximately one second in a loop with 0-5
seconds delay.
NOTE: The current version of the script stores the output in
$HOME/power-monitoring-db— so make sure to modify this is you want it somewhere else.
You may need to run modprobe intel_rapl_msr to enable RAPL readings. Also,
make sure that you have an awk-version that supports asort (e.g. gawk over
mawk).
I recommend reading up on Intel RAPL for how to interpret the results, e.g. at:
- Issue 116 in the Scaphandre repo.
- “RAPL in Action: Experiences in Using RAPL for Power Measurements,” (K. N. Khan, M. Hirki, T. Niemi, J. K. Nurminen, and Z. Ou, ACM Trans. Model. Perform. Eval. Comput. Syst., vol. 3, no. 2, pp. 1–26, Apr. 2018, doi: 10.1145/3177754).
Much like the power monitoring, the sysstat monitoring runs as a system service. When active, it will read system resource usage during 60 seconds once every half-hour.