Open tilsche opened 6 years ago
Looking at it a bit more in detail, these seem to be the RobustConnection._closing
futures. It seems that the error is delayed by the GC, which is why it happened for me in-between and at the end.
Would it make sense to cancel _closing
before assigning a new future because we don't even know if someone will ever wait for this future?
Same here. I see such error every rabbitmq restart. And it's not delayed. I don't use set_exception_handler but handle errors in place (with monitoring and checking task.exception()
).
https://github.com/mosquito/aio-pika/issues/112
Offers an intermediate solution until robust_connection() is more suitable to your requirements.
I am trying to setup a robust connection like this:
Basically this works. During a reconnect, there are some reported Exceptions such as
and
But if the connection resumes, and I stop the program, then at the very end I get a lot of those
This makes me worry that there remains something unclear from the reconnection that could gather over time. Is there something to be done differently in the code with respect to error handling and cleanup?