I have a seen a recipe set which got itself into a supposedly impossible state: two recipes, one is Aborted due to external watchdog (it started running) while the other is Scheduled. All recipes in a recipe set are supposed to be started at the same time which is why this should be impossible.

A symptom is that beaker-watchdog loops repeatedly trying to abort the already-Aborted recipe (whereas the Scheduled recipe has no watchdog yet). So as part of this fix we should make sure that recipes which are already Aborted are not returned in the active watchdog list.