pulpcore.tasking.worker:ERROR: A worker with name 1@pulp-worker-2.foremancontainer.linux.schnell.er already exists in the database.
This is your issue, there is a database entry for the worker (probably from the before-the-reboot) and thus the worker now refuses to start.
In theory (!) this should not happen, but you running into this clearly shows it does 
Some time ago we implemented ensure apps are started after databases, and stopped before by evgeni 路 Pull Request #291 路 theforeman/foremanctl 路 GitHub that is supposed to order the services in a way that during shutdown the pulp workers have enough time to clean up the DB before the DB is shut down itself. This clearly didn鈥檛 work for you.
Could you provide the logs (from journalctl) of the pulpcore workers and the PostgreSQL for the time between you typed reboot and the machine was off?