I am working on the problem that Loic dubbed "server reduce backlog",
which is described at the following URL: http://ur1.ca/4bk

As of r4401, I have produced test73 in test-pokeravatar.py.in that I
believe shows the problem is not occurring at the PokerAvatar level, at
least not in a way that I can see.

I therefore cannot figure out where in the server stack the packets are
being queued.  I had assumed the problem was occurring because
PokerAvatar.queuePackets() was called, and then the packets were being
left in that queue indefinitely until the client reconnected.  test73
shows that clearly isn't the case, so there is more research to be done
on this question.

One thought might be that PokerAvatar.queuePackets() is not being called
at all when the sitOut in question occurs, and therefore all packets
_are_ being sent. Then, would they get queued by Twisted? By
pokerservice? By the network itself?

Any thoughts that folks have would be appreciated.  I assume this task
got created for the server because the behavior was witnessed in
jspoker.  Can someone rerun that and get a better sense from the client
side what conditions are created that cause this influx of packets?


I am ending my research now for the night and will check on #pokersource
with everyone tomorrow.
-- 

   -- bkuhn

_______________________________________________
Pokersource-users mailing list
[email protected]
https://mail.gna.org/listinfo/pokersource-users

Reply via email to