> On Sun, 9 Nov 2008 00:22:11 +0000, > Matthew Toseland <[EMAIL PROTECTED]> wrote: > > Freenet 0.7 build 1173 is now available.
It is still using much more of (my 1.2GHz) cpu than I would like. It regularly hovers at 100% for a few minutes, then calms down, then repeats. And, probably not unrelated, it is still producing LOTS of PacketSender ERRORs in the freenet logfiles. (I am forced to disable logging, otherwise it could /easily/ grow out of control.) For example, just now I enabled logging for /less/ than 2 minutes, and it grew to over 2MB, *ENTIRELY* composed of freenet.node.KeyTracker, PacketSender thread, Packet bla sent over blams ago still not acked ERRORs. It has been brought up on IRC a number of times today, most notably by mkan--but I forget if a reason/solution was given. It may not be isp throttling as I originally guessed, since mkan lives in a different part of the world, and is experiencing the same thing--as I think many others are as well. In this last 2min test of mine, all of the errors came from my 1173 opennet peers. Only one of them was listed as "BACKED OFF" (I was/am connected to him for 4h49m)--the other two were (and still are) "CONNECTED" (for 48m and 1h21m). They have varying overload probability percentages, 14% 16% 42% respectively. They have varying congestion control rates, 1MB/s 1MB/s 17Bytes/s respectively. They have varying RTT times, 263ms 138233ms 71006ms. They have varying overloaded percentages, the "backed off" one was "99.8% 7974/10800/AcceptedTimeout" but the other two were 3% and 28%, both with 0/1/ForwardRejectedOverload. _______________________________________________ Support mailing list Support@freenetproject.org http://news.gmane.org/gmane.network.freenet.support Unsubscribe at http://emu.freenetproject.org/cgi-bin/mailman/listinfo/support Or mailto:[EMAIL PROTECTED]
