AttributeError: 'NoneType' object has no attribute 'recv'

[weswitt]

I'm running version 1.4 (latest). Code is running on Ubuntu 18.04 in a VirtualBox VM.

I'm seeing the following exception and it happens very reliably. The app does not run very long before this throws.

Exception in thread Thread-1:
Traceback (most recent call last):
  File "/usr/lib/python2.7/threading.py", line 801, in __bootstrap_inner
    self.run()
  File "/usr/lib/python2.7/threading.py", line 754, in run
    self.__target(*self.__args, **self.__kwargs)
  File "/usr/local/lib/python2.7/dist-packages/paho/mqtt/client.py", line 2913, in _thread_main
    self.loop_forever(retry_first_connection=True)
  File "/usr/local/lib/python2.7/dist-packages/paho/mqtt/client.py", line 1578, in loop_forever
    rc = self.loop(timeout, max_packets)
  File "/usr/local/lib/python2.7/dist-packages/paho/mqtt/client.py", line 1072, in loop
    rc = self.loop_read(max_packets)
  File "/usr/local/lib/python2.7/dist-packages/paho/mqtt/client.py", line 1374, in loop_read
    rc = self._packet_read()
  File "/usr/local/lib/python2.7/dist-packages/paho/mqtt/client.py", line 2030, in _packet_read
    byte = self._sock_recv(1)
  File "/usr/local/lib/python2.7/dist-packages/paho/mqtt/client.py", line 607, in _sock_recv
    return self._sock.recv(bufsize)
AttributeError: 'NoneType' object has no attribute 'recv'

[weswitt]

Dropping back to 1.3.1 and this problem does not happen. Seems like a bug introduced in 1.4

[sousacanfly]

I can confirm, the same's happening here.
Using latest version too.

[weswitt]

With some more debugging I worked around this. The issue was caused by the fact that I was using the multi-threaded loop_forever. What happens is that the socket connection is dropped and then my main thread called reconnect. This causes the socket object to be set to None (null) and then you see this error. I changed my code to use the single threaded loop and it no longer happens. That being said, this is a bug in the paho mqtt code.

Also, my previous statement about 1.3.1 fixing it was false. It also happens with that version.

[PierreF]

loop_forever could not and should not be run in multiple thread. We should either add some note in the documentation and/or some kind of assert in the code that reject running loop_forever twice.

[weswitt]

You misread my comment. I did and would not run loop_forever twice.

[PierreF]

What do you mean by the multi-threaded loop_forever ?

[kitten77]

@PierreF

ERROR asyncio    Exception in callback None()
handle: <Handle cancelled>
Traceback (most recent call last):
  File "/usr/local/lib/python3.7/asyncio/events.py", line 88, in _run
    self._context.run(self._callback, *self._args)
  File "/home/muad/workbench/tonder/irc/plugins/mqtt.py", line 30, in cb
    client.loop_read()
  File "/home/muad/venv/tonder3.7/lib/python3.7/site-packages/paho/mqtt/client.py", line 1374, in loop_read
    rc = self._packet_read()
  File "/home/muad/venv/tonder3.7/lib/python3.7/site-packages/paho/mqtt/client.py", line 2030, in _packet_read
    byte = self._sock_recv(1)
  File "/home/muad/venv/tonder3.7/lib/python3.7/site-packages/paho/mqtt/client.py", line 607, in _sock_recv
    return self._sock.recv(bufsize)
AttributeError: 'NoneType' object has no attribute 'recv'

Any ideas on what is causing this?

[emha69]

This seems to be caused if the client.loop(n>1), and if the client.loop() is not used immediately after client.publish().
For example, when the client.loop() is used in the main loop e.g. for waiting for messages or something like that.

[bggardner]

I had this error when using loop_start() and calling connect() more than once, not realizing the client auto-reconnects. So, not calling connect() again was my solution.

[mxmaxime]

Got the same issue here, sometimes it works sometimes not... After a bit of research I found that I had loop_start() and after a loop_forever(), removing one of them worked.

I've searched a bit in the library, I don't understand how self._sock could be None, it's only set to None in the method _sock_close(). I didn't searched any further, but in my understanding:

• checks on self._sock is not None are performed before calling self._sock_recv, but in between another thread set it to None by calling _sock_close().

[russell-sealand]

I fixed the problem by switching to use client.loop_start() instead of client.loop_forever()