Mentat

The DB upgrade procedure, as described in #7555, allows for reboot in step 9, if the circumstances require it. During execution on mentat-hub, after the reboot, the Mentat's import pipeline (the daemons) were running. No data was actually imported as the warden-filer was disabled at systemctl level (and the issue described in #7121 also surfaced).
In step 1, Mentat is stopped and disabled using mentat-controller.py, so there seems to be some inconsistency. Should those two steps actually be enough to prevent the pipeline from running?
Setting low priority as this is actually a very rare scenario.