This MR still needs small adjustment: dummy database is now generated separately from the main LAVA CI pipeline.
