Lots of "Network Core Overloaded"?

[whizzmoraza]

I was seeing what I considered to be quite a few "Network Core overloaded" messages in Shareaza's System log. As a sanity check, I enabled the debug log at level 3. I let the log file build up over 24 hours. During this 24 hours, I had ~10 small downloads in queue. My upload queue was nearly always full (natch). Here's what I came up with:

24 hours worth of log file entries at log level 3:
Total entries: 68917
"network core overloaded": 7876 (~11.43% of total)
"malformatted query packet": 4421 (~6.42% of total - all from G1 ultrapeers, mostly Limewire 5.4.x)

My pf is set up correctly, and the connection test comes back green for TCP and UDP. I am running the current nightly (20100207). [Side note: The version displayed by Shareaza is "r8462 20100207" even though the DL link showed r8465.]

My Windows EventLog is free of any "event id 4662" messages.

Is it normal to see this many "network core overloaded" warnings?

-Whizzmo

[raspopov]

[whizzmoraza]

[whizzmoraza]

[mojo85]

One thing you can test out is to increase Shareaza's timeout periods in the Advanced Settings menu for certain types of connections and observe how Shareaza responds. I also noticed the same thing when I turn on my QOS on the Router, and less of them when the QOS is off.

[ailurophobe]

If Shareaza interpreted failed connection attempts as "network core overload" that would explain why I am occasionally seeing all those as well. Is there an advanced option for telling Shareaza it is behind an IPfilter? Or a way to make it not count failed retries just first attempts. That would reduce the effect of IP filtering without preventing the detection of real overload.

[raspopov]