Core & Internals: fix bulk rse loading

- settings_by_id will not have any entry for RSEs which don't have
attributes from RSE_SETTINGS set. This resulted in an attribute error.

- name loading was getting AttributeError when some RSEs already had
their name loaded.

- remove code for non-temp-table workflow

lib/rucio/core/rse.py

@@ -159,20 +159,6 @@ def bulk_load(rse_id_to_data: "dict[str, RseData]", load_name=False, load_column
         Given a dict of RseData objects indexed by rse_id, ensure that the desired fields are initialised
         in all objects from the input.
         """
-        use_temp_tables = config_get_bool('core', 'use_temp_tables', default=True, session=session)
-        if not use_temp_tables or len(rse_id_to_data) < 4:
-            for rse_data in rse_id_to_data.values():
-                rse_data.ensure_loaded(
-                    load_name=load_name,
-                    load_columns=load_columns,
-                    load_attributes=load_attributes,
-                    load_info=load_info,
-                    load_usage=load_usage,
-                    load_limits=load_limits,
-                    session=session
-                )
-            return
-
         rse_ids_to_load = set()
         for rse_id, rse_data in rse_id_to_data.items():
             anything_to_load = any((
@@ -224,7 +210,7 @@ def bulk_load(rse_id_to_data: "dict[str, RseData]", load_name=False, load_column
                                                                   session=session))
             for rse_id, db_rse in db_rses_by_id.items():
                 rse_data = rse_id_to_data[rse_id]
-                settings = rse_data._attributes if rse_data._attributes is not None else settings_by_id[rse_id]
+                settings = rse_data._attributes if rse_data._attributes is not None else settings_by_id.get(rse_id, {})
                 columns = _format_get_rse(db_rse=db_rse, rse_attributes=settings, session=session)
                 rse_data._columns = columns
                 rse_data._name = columns['rse']
@@ -281,8 +267,8 @@ def bulk_load(rse_id_to_data: "dict[str, RseData]", load_name=False, load_column
         if load_name:
             # The name could have been loaded already (when loading columns or info). Skip loading if it's known.
             if not load_columns and not load_info:
-                for rse_id, rse_data in rse_id_to_data.items():
-                    rse_data._name = db_rses_by_id[rse_id].rse
+                for rse_id in rse_ids_to_load:
+                    rse_id_to_data[rse_id]._name = db_rses_by_id[rse_id].rse
 
 
 class RseCollection(Generic[T]):