On Jun 14, 2012, at 1:58 AM, Diana Marosin wrote:

> Hello guys,
> 
> I am trying to run spanning tree in POX (./pox py forwarding.l2_learning 
> openflow.discovery openflow.spanning_tree) and I receive some errors from the 
> python library (see attachment) regarding threading.py and socket.py and from 
> the pox library, recoco.py and of_01.py
> 
> Can you have a short look and tell me if you had those before and if you have 
> any idea how they appeared and how to fix them?

Hi, Diana.

I think this bug has been mentioned on the pox-dev mailing list recently.  I 
think this is also bug #20 in the POX issue tracker, and possibly bug #18 as 
well (though I'm not sure about the latter one).

Issue #20 can be found at: https://github.com/noxrepo/pox/pull/20

This has a pull request with code that addresses the problem, but I think the 
problem is actually a bit deeper and that the main patch there addresses the 
symptom and not the root issue.

I *think* the underlying problem here is that Connection.disconnect() has the 
meanings for its "hard" parameter backwards, and that hard=False should not 
have a call to close() at all.  But basically that part of the code is just 
sort of confusing and messy.  I have attempted to clean this up, and provided a 
diff later in the issue #20 discussion.

You can also find my patch at https://gist.github.com/2929386.  I'd love 
feedback -- I haven't been able to reproduce the problem myself. 

-- Murphy

P.S., I've copied the pox-dev list, which probably makes more sense for further 
discussion, but it also might make sense to discuss it on the github issue.

Reply via email to