bleak
bleak copied to clipboard
hbldh
"bleak.exc.BleakError: This service is already present in this BleakGATTServiceCollection!" when reconnecting a device
- bleak version: 0.9.1
- Python version: 3.7
- BlueZ version (
bluetoothctl -v) in case of Linux: 5.50
Description
For very first time the central code for scanning BLE Devices worked and connected with uart device. But after connection I was not able to disconnect it.
What I Did
I had scanned BLE Devices and make connection. It worked and shown
disconnected, scanning connected
Then I had stopped peripheral service. After service was stopped I had ran scanner class again then i was showing below error.
disconnected, scanning
Traceback (most recent call last):
File "echoClient.py", line 31, in <module>
ble.connect(advertisement)
File "/usr/local/lib/python3.7/dist-packages/adafruit_ble/__init__.py", line 305, in connect
connection = self._adapter.connect(advertisement.address, timeout=timeout)
File "/usr/local/lib/python3.7/dist-packages/_bleio/adapter_.py", line 373, in connect
return self.await_bleak(self._connect_async(address, timeout=timeout))
File "/usr/local/lib/python3.7/dist-packages/_bleio/adapter_.py", line 134, in await_bleak
return future.result(timeout)
File "/usr/lib/python3.7/concurrent/futures/_base.py", line 432, in result
return self.__get_result()
File "/usr/lib/python3.7/concurrent/futures/_base.py", line 384, in __get_result
raise self._exception
File "/usr/local/lib/python3.7/dist-packages/_bleio/adapter_.py", line 384, in _connect_async
await client.connect(timeout=timeout)
File "/usr/local/lib/python3.7/dist-packages/bleak/backends/bluezdbus/client.py", line 157, in connect
await self.get_services()
File "/usr/local/lib/python3.7/dist-packages/bleak/backends/bluezdbus/client.py", line 405, in get_services
BleakGATTServiceBlueZDBus(service, object_path)
File "/usr/local/lib/python3.7/dist-packages/bleak/backends/service.py", line 116, in add_service
"This service is already present in this BleakGATTServiceCollection!"
bleak.exc.BleakError: This service is already present in this BleakGATTServiceCollection!
Do you have two services with the same uuid on the peripheral? In that case this is related to #362.
No I have ran only one service with UART_SERVICE_UUID = '6e400001-b5a3-f393-e0a9-e50e24dcca9e' UART_RX_CHARACTERISTIC_UUID = '6e400002-b5a3-f393-e0a9-e50e24dcca9e' UART_TX_CHARACTERISTIC_UUID = '6e400003-b5a3-f393-e0a9-e50e24dcca9e' There is no other service with same UUID
I'm seeing something perhaps related, when I've got a device that drops connectivity (powered off) and then try to reconnect programmatically. If you think this is a different issue, more than willing to split it off. I'm still diagnosing and trying to discover "why".
No minimal, reproducible case yet.
bluetoothctl: 5.50 PRETTY_NAME="Raspbian GNU/Linux 10 (buster)" Linux pi-walnut 5.10.52-v7+ #1440 SMP Tue Jul 27 09:54:13 BST 2021 armv7l GNU/Linux Python 3.9.5 bleak 0.12.1
Even if I try an explicit disconnect, the problem still seems to be present.
In the disconnect callback
if not client.willful_disconnect:
asyncio.get_event_loop().create_task(self._reconnect())
async def _reconnect(self):
"""
Will try immediately, then 1, 2, 3, ..., 10, 10, ... seconds later
"""
await self.disconnect()
class_name = type(self).__name__
if self._logging_reconnect:
logger.info(
f"Will try reconnecting to {class_name} at {self.address} "
f"after waiting {self._reconnect_delay} seconds.")
await asyncio.sleep(self._reconnect_delay)
await self.connect()
if self.is_connected:
self._reconnect_delay = 0
self._logging_reconnect = True
else:
# code to try again in a bit
Traceback (most recent call last):
File "/home/ble-remote/devel/pyDE1/src/pyDE1/scale/scale.py", line 478, in _reconnect
await self.connect()
File "/home/ble-remote/devel/pyDE1/src/pyDE1/scale/scale.py", line 172, in connect
await asyncio.gather(self._event_connectivity.publish(
File "/home/ble-remote/devel/pyDE1/src/pyDE1/bleak_client_wrapper.py", line 88, in connect
retval = await super(BleakClientWrapped, self).connect(**kwargs)
File "/home/ble-remote/venv/pyDE1_3.9/lib/python3.9/site-packages/bleak/backends/bluezdbus/client.py", line 258, in connect
self.services.add_service(BleakGATTServiceBlueZDBus(obj, path))
File "/home/ble-remote/venv/pyDE1_3.9/lib/python3.9/site-packages/bleak/backends/service.py", line 125, in add_service
raise BleakError(
bleak.exc.BleakError: This service is already present in this BleakGATTServiceCollection!
The wrapper tries to manage clean-up atexit, marginally successfully (mainly around if there is a running loop and things complete before Python shuts it down).
async def connect(self, **kwargs) -> bool:
atexit.register(self.sync_disconnect)
retval = await super(BleakClientWrapped, self).connect(**kwargs)
if retval:
self._willful_disconnect = False
return retval
async def disconnect(self) -> bool:
self._willful_disconnect = True
retval = await super(BleakClientWrapped, self).disconnect()
self._willful_disconnect = False
if retval:
logger.debug("Unregistering atexit disconnect "
f"{self.name} at {self.address}")
atexit.unregister(self.sync_disconnect)
return retval
On powering back on the Skale, it does show up in bluetoothctl
[CHG] Device CF:75:75:90:8C:E3 ServicesResolved: no
[CHG] Device CF:75:75:90:8C:E3 Connected: no
[DE1]# info CF:75:75:90:8C:E3
Device CF:75:75:90:8C:E3 (random)
Name: Skale
Alias: Skale
Paired: no
Trusted: no
Blocked: no
Connected: no
LegacyPairing: no
UUID: Generic Access Profile (00001800-0000-1000-8000-00805f9b34fb)
UUID: Generic Attribute Profile (00001801-0000-1000-8000-00805f9b34fb)
UUID: Device Information (0000180a-0000-1000-8000-00805f9b34fb)
UUID: Battery Service (0000180f-0000-1000-8000-00805f9b34fb)
UUID: Unknown (0000ff08-0000-1000-8000-00805f9b34fb)
UUID: Vendor specific (0f050001-3225-44b1-b97d-d3274acb29de)
ManufacturerData Key: 0x81ef
ManufacturerData Value:
fe ff ff ff
Without deep knowledge of what is supposed to be happening, it is my guess that a previous service discovery is not being cleared before adding in the re-discovered services. Breakpoint at line 325 in service.py
To reproduce:
- Configure the BLE_ID
- Turn the device on and enable its advertisements
- Run the program
- Observe that the device connects
- Turn the device off
- Note the logs indicating that
reconnect()was called - Note the (unretrieved) error (as the task result wasn't accessed)
bleak.exc.BleakError: This characteristic is already present in this BleakGATTServiceCollection!
import asyncio
import logging
from bleak import BleakClient
format_string = "%(asctime)s %(levelname)s %(name)s: %(message)s"
logging.basicConfig(level=logging.DEBUG,
format=format_string,
)
logger = logging.getLogger('run')
def disconnect_handler(c: BleakClient):
logger.info("Disconnect handler")
asyncio.get_running_loop().create_task(reconnect(c))
async def reconnect(c: BleakClient):
logger.info("Reconnect")
await c.connect()
if not c.is_connected:
# lazily call recursively, as shouldn't be more than a few
await asyncio.sleep(10)
await reconnect(c)
async def run():
BLE_ID = 'CF:75:75:90:8C:E3'
client = BleakClient(BLE_ID)
client.set_disconnected_callback(disconnect_handler)
await client.connect()
logger.info("Connected")
await asyncio.sleep(90)
logger.info("90 seconds, disconnecting")
await client.disconnect()
if __name__ == '__main__':
loop = asyncio.get_event_loop()
loop.set_debug(True)
loop.run_until_complete(run())
Ugly fix, but this at least appears to be functional for this specific host and implementation
Use "private access" with c.services = BleakGATTServiceCollection()
async def reconnect(c: BleakClient):
logger.info("Reconnect")
c.services = BleakGATTServiceCollection()
await c.connect()
if not c.is_connected:
# lazily call recursively, as shouldn't be more than a few
await asyncio.sleep(10)
await reconnect(c)
For those seeing this issue when reconnecting after a disconnect, I think https://github.com/hbldh/bleak/pull/621#discussion_r687152961 will fix the issue (it essentially does this: https://github.com/hbldh/bleak/issues/376#issuecomment-898008057).
If anyone is seeing this error at a different time, it would be interesting/helpful if you could confirm the sequence of events I describe in https://github.com/hbldh/bleak/pull/625#issuecomment-902868577.