Skip to main content

Configuration via config.ini

How configuration works across Pioreactors​

Configuration of the Pioreactors is through configuration files, prefixed by .ini. There are two types of .ini files: a shared config.ini, which is the same across all units, and a worker-specific unit_config.ini file, which can have settings that overwrite those found in the shared config.ini. See image below:

This overriding lets you customize specific behavior per Pioreactor, without having to duplicate lots of configuration over and over again.

Editing the configuration files from the web interface​

In the web interface, pioreactor.local, the sidebar contains a link to Configuration. From there, any .ini file can be edited. Clicking [Save] will save the new configuration and will deploy the new configuration to the proper Pioreactor unit(s).

Editing the configuration files from the command line​

All the .ini files come from the leader unit. They are stored in /home/pioreactor/.pioreactor - one config.ini and a config_<unitName>.ini for each Pioreactor unit. The command pios sync-configs will deploy the .ini files to their correct unit. Editing the .ini files on the leader is correct workflow, as any edits on the worker units will be overwritten on the next pios sync-configs.

One can programmatically, or by hand, edit the config_<unitName>.ini files to create varying parameters in an experiment. Running pios sync-configs will deploy the config files.

Editing the config.ini before booting​

Sometimes it's desirable to make changes to the config.ini without having to boot the Pioreactor first. You can do this by plugging the SD card into your computer, and adding a config.ini to the boot folder on it. This config.ini doesn't need to have all the fields, only the fields you wish to edit. At start up, the Pioreactor will merge this /boot/config.ini into the main config.ini (and delete it).

Example config.ini​

If you need a reference config.ini, or to replace parts of yours, here's our default one ships with the Raspberry Pi image. To use it, you'll need to make some changes, specifically to `cluster.leader` and `inventory`.