This MR changes scope_id from 1 to 3. There should never be a call to
option_db_record_reservation() with something other than
host because options that belong to reservations always have
host scope. If the reviewer so desires,
host could also be hardcoded into the function.
Only Postgres tests are impacted by this bug, but this MR changes MySQL so that the tests properly use reservation option-data as intended and not subnet option-data.
Should fix the following tests:
The database results should be the same as after a