Closed mffrench closed 7 years ago
Hi, apologies for the delay in this one. This would have happened when the client running Python 3.4 attempted to reconnect where there were some bugs with the task cancellation.
I've added a couple of fixes which might help with reaching this condition and released them under version v0.4.0.
As you mention, I'm not sure of the mileage we will get with asyncio in Python 3.4 and threadsafety so would recommend possibly upgrading to 3.5 and using the run_coroutine_threadsafe APIs, though native pykka driver without asyncio could be interesting as well...
Hi,
I'm currently facing an issue with asyncio loop after more than 48h run (so not easy to reproduce).
This asyncio problem (some loop getters problem I need to troubleshoot deeper) impact the asyncio-nats code as shown bellow :
Maybe is there some protection code to add here ? If you have any tips to help me troubleshoot deeper let me now anyway...
Here's the code where I'm using asyncio-nats : https://github.com/echinopsii/net.echinopsii.ariane.community.cli.python3/blob/master/ariane_clip3/natsd/driver.py
NOTE : I'm using actor pattern (pykka) with asyncio-nats which may not be a great fit... Probably should I use deeper functions (IE : function implementing NATS protocol only) in asyncio-nats to avoid asyncio logic ?
Full stack trace :
Thank you for your help :)