Skip to main content
Version: v3

Experimental Mode

Before making certain features generally available, odo might expose them as experimental first.

This gives you an opportunity to try out our newest features sooner. In return, your feedback helps us make sure that our new features are reliable and useful, and we appreciate any and all feedback you want to provide.

If a feature is labeled as Experimental, this specifically means:

  • The feature is in early or intermediate stages of development and your feedback is highly appreciated.
  • The feature may have known and undiscovered bugs.
  • The feature may be changed, deprecated or removed altogether.
  • The documentation for the feature can be incomplete, missing or may contain errors.
  • The feature will have to be deliberately enabled.

Enabling the experimental mode

Experimental mode is currently opt-in. You can enable it by setting the ODO_EXPERIMENTAL_MODE environment variable to true prior to running odo. Doing so unlocks all experimental commands and flags.


$ ODO_EXPERIMENTAL_MODE=true odo dev --run-on=some-platform

Experimental mode enabled for flag: --run-on. Use at your own risk. More details on
- Forwarding from -> 8080

Running odo with an experimental command or flag without enabling the experimental mode returns an error.

$ odo dev --run-on=some-platform
✗ unknown flag: --run-on

List of experimental features

Generic --run-on flag

This is a generic flag that allows running odo on any supported platform (other than the default Kubernetes or OpenShift cluster mode).

The supported platforms are cluster and podman.

By default, if you do not use the --run-on flag, or if you do not activate the experimental mode, the cluster platform is used.

The cluster platform uses the current Kubernetes or OpenShift cluster.

The podman platform uses the local installation of podman. It relies on the podman binary to be installed on your system.

These commands support the --run-on flag:

  • odo dev