Improve D-Bus timeout error handling (#5664)

* Improve D-Bus timeout error handling

Typically D-Bus timeouts are related to systemd activation timing out
after 25s. The current dbus-fast timeout of 10s is well below that
so we never get the actual D-Bus error. This increases the dbus-fast
timeout to 30s, which will make sure we wait long enought to get the
actual D-Bus error from the broker.

Note that this should not slow down a typical system, since we tried
three times each waiting for 10s. With the new error handling typically
we'll end up waiting 25s and then receive the actual D-Bus error. There
is no point in waiting for multiple D-Bus/systemd caused timeouts.

* Create D-Bus TimedOut exception

supervisor/dbus/udisks2/__init__.py

@@ -66,8 +66,8 @@ async def connect(self, bus: MessageBus):
         try:
             await super().connect(bus)
             await self.udisks2_object_manager.connect(bus)
-        except DBusError:
-            _LOGGER.warning("Can't connect to udisks2")
+        except DBusError as err:
+            _LOGGER.critical("Can't connect to udisks2: %s", err)
         except (DBusServiceUnkownError, DBusInterfaceError):
             _LOGGER.warning(
                 "No udisks2 support on the host. Host control has been disabled."

supervisor/exceptions.py

@@ -403,7 +403,11 @@ class DBusParseError(DBusError):
 
 
 class DBusTimeoutError(DBusError):
-    """D-Bus call timed out."""
+    """D-Bus call timeout."""
+
+
+class DBusTimedOutError(DBusError):
+    """D-Bus call timed out (typically when systemd D-Bus service activation fail)."""
 
 
 class DBusNoReplyError(DBusError):

supervisor/os/data_disk.py

@@ -189,12 +189,13 @@ async def load(self) -> None:
             await self.sys_dbus.agent.datadisk.reload_device()
 
         # Register for signals on devices added/removed
-        self.sys_dbus.udisks2.udisks2_object_manager.dbus.object_manager.on_interfaces_added(
-            self._udisks2_interface_added
-        )
-        self.sys_dbus.udisks2.udisks2_object_manager.dbus.object_manager.on_interfaces_removed(
-            self._udisks2_interface_removed
-        )
+        if self.sys_dbus.udisks2.is_connected:
+            self.sys_dbus.udisks2.udisks2_object_manager.dbus.object_manager.on_interfaces_added(
+                self._udisks2_interface_added
+            )
+            self.sys_dbus.udisks2.udisks2_object_manager.dbus.object_manager.on_interfaces_removed(
+                self._udisks2_interface_removed
+            )
 
     @Job(
         name="data_disk_migrate",

supervisor/utils/dbus.py

@@ -31,6 +31,7 @@
     DBusObjectError,
     DBusParseError,
     DBusServiceUnkownError,
+    DBusTimedOutError,
     DBusTimeoutError,
     HassioNotSupportedError,
 )
@@ -87,6 +88,8 @@ def from_dbus_error(err: DBusFastDBusError) -> HassioNotSupportedError | DBusErr
             return DBusNotConnectedError(err.text)
         if err.type == ErrorType.TIMEOUT:
             return DBusTimeoutError(err.text)
+        if err.type == ErrorType.TIMED_OUT:
+            return DBusTimedOutError(err.text)
         if err.type == ErrorType.NO_REPLY:
             return DBusNoReplyError(err.text)
         return DBusFatalError(err.text, type_=err.type)
@@ -136,15 +139,21 @@ async def introspect(self) -> Node:
         for _ in range(3):
             try:
                 return await self._bus.introspect(
-                    self.bus_name, self.object_path, timeout=10
+                    self.bus_name, self.object_path, timeout=30
                 )
             except InvalidIntrospectionError as err:
                 raise DBusParseError(
                     f"Can't parse introspect data: {err}", _LOGGER.error
                 ) from err
             except DBusFastDBusError as err:
                 raise DBus.from_dbus_error(err) from None
-            except (EOFError, TimeoutError):
+            except TimeoutError:
+                # The systemd D-Bus activate service has a timeout of 25s, which will raise. We should
+                # not end up here unless the D-Bus broker is majorly overwhelmed.
+                _LOGGER.critical(
+                    "Timeout connecting to %s - %s", self.bus_name, self.object_path
+                )
+            except EOFError:
                 _LOGGER.warning(
                     "Busy system at %s - %s", self.bus_name, self.object_path
                 )