I noticed, that there is a (new) installation documentation using docker. I thus wanted to ask, what the recommended installation approach (currently/in future) is? Using systemd service, or docker?
Can’t talk about what’s officially recommend but for most things I run I prefer docker as it makes updating a lot easier and less likely to run into dependency issues and you have UIs like portainer that make it pretty straight forward to manage.
Although if you are planning to maintain/compile your own releases or even build a ready to go OS image that you for example just flash on the SD card of a Rapsberry there isn’t really much of a need to go with docker.
2 Likes
We introduced Docker, because it makes testing OpenEMS a lot easier.
At the same time most OpenEMS distributions out there are customized and run on small IoT devices (like a Raspberry Pi) and then building/compiling manually and running as native systemd services is favourable.
1 Like
Thx for the reply. If I understand it right, having everything on one common LAN, my thoughts about a good setup (productive for a small village, no intention of customizing or maintaining an own release) are:
- Backend + UI in one shared debian VM on PVE → Installation via Docker
- Edge for devices via ethernet in one debian VM on PVE → Installation via Docker
- (if needed) Edge for devices via serial connection or other non-ethernet connection on IoT devices (like Raspberry Pi) → native installation