AWS_ERROR_MQTT_NOT_CONNECTED

[1960pghpirates]

Confirm by changing [ ] to [x] below:

• [x ] I've searched for previous similar issues and didn't find any solution

Platform/OS/Device
Raspberry Pi Zero WH

Describe the question

I recently upgraded from v1 to v2 of the AWS IoT Python SDK. The client ran without error for a couple of weeks, but now a new error has stopped it:

Failed to publish update request.
Exiting sample due to exception.
Traceback (most recent call last):

File "sgp40_simpletest-new.py", line 227, in on_publish_update_shadow
future.result()
File "/usr/lib/python3.7/concurrent/futures/_base.py", line 425, in result
return self.__get_result()
File "/usr/lib/python3.7/concurrent/futures/_base.py", line 384, in __get_result
raise self._exception
File "/home/pi/.local/lib/python3.7/site-packages/awscrt/mqtt.py", line 680, in publish
packet_id = _awscrt.mqtt_client_connection_publish(self._binding, topic, payload, qos.value, retain, puback)
RuntimeError: 5131 (AWS_ERROR_MQTT_NOT_CONNECTED): The requested operation is invalid as the connection is not open.

The only thing I can think of that might cause this is WiFi connection drops, but from what I read online, the client is supposed to be able to recover from periodic connection issues. Is that true? I never had this problem with v1 of the Python SDK.

[jmklix]

Did you change any of the connection builder arguments from the default? Specifically the keep_alive_sec

Can you attach logs as well as include a code snippet that reproduces this?

Yes the client should recover from temporary disconnects less than 1.5 times the keep_alive_sec

keep_alive_secs (int): The keep alive value, in seconds, to send in CONNECT packet.
A PING will automatically be sent at this interval. The server will assume the connection is lost if no PING is received after 1.5X this value. Default is 1200sec (20 minutes). This duration must be longer than ping_timeout_ms.

[1960pghpirates]

What is the name of the log file that would have the information you need? Here is the code snippet, keep alive parameter is 30 seconds. I do not think I changed it from the default example provided for Python SDK v2.

if args.use_websocket == True:
credentials_provider = auth.AwsCredentialsProvider.new_default_chain(client_bootstrap)
mqtt_connection = mqtt_connection_builder.websockets_with_default_aws_signing(
endpoint=args.endpoint,
client_bootstrap=client_bootstrap,
region=args.signing_region,
credentials_provider=credentials_provider,
http_proxy_options=proxy_options,
ca_filepath=args.root_ca,
client_id=args.client_id,
clean_session=True,
keep_alive_secs=30)

else:
    mqtt_connection = mqtt_connection_builder.mtls_from_path(
        endpoint=args.endpoint,
        cert_filepath=args.cert,
        pri_key_filepath=args.key,
        client_bootstrap=client_bootstrap,
        ca_filepath=args.root_ca,
        client_id=args.client_id,
        clean_session=True,
        keep_alive_secs=30,
        http_proxy_options=proxy_options)

print("Connecting to {} with client ID '{}'...".format(
    args.endpoint, args.client_id))

connected_future = mqtt_connection.connect()

[jmklix]

the logs can be output to the console with the following command:
io.init_logging(io.LogLevel.Trace, 'stderr')
Also you can pass an optional parameter to the samples to change the verbosity: --verbosity Trace

[1960pghpirates]

Ok, the client is running with logging enabled at the Trace level of verbosity. It may take some time for the error to reoccur. Is 30 seconds an acceptable value for the keep alive parameter? That means than any WiFi disruption of 45 seconds or more will cause a loss of MQTT connection?

[jmklix]

It will ultimately depend on what requirements you need for your application. The default is 20 min, and the samples use 30 seconds. If you expect WiFi disruptions of 45 seconds or more to be common, you might want a longer keep-alive time. For testing and reproducing shorter is probably easier, so you can reproduce the error faster.