If the broker gets stopped while starting up (as part of a bundle refresh during Fabric container startup) it may raise a BrokerStoppedException and it may fail to properly shut down the Scheduler KahaDB instance.

Upon broker restart (again, as part of bundle refresh) the broker will fail start up fully as the Scheduler KahaDB instance is still locked.
The broker will raise

Database data/mybroker/mybroker/scheduler/lock is locked by another server. This broker is now in slave mode waiting a lock to be acquired

and will fail to start up fully