On Thu, Aug 07, 2003 at 11:44:23PM +0100, Gordan wrote: > On Thursday 07 August 2003 23:02, Greg Wooledge wrote: > > Gordan ([EMAIL PROTECTED]) wrote: > > > Has anything happened very recently (since 6137+) that would explain a > > > drastic reduction in bandwidth consumption of my node? I have not changed > > > the limits, but the used bandwidth is signifficantly reduced. The limits > > > seem to be reported same as before in the logs, too. I'd like to know for > > > sure there is a reason for this before I start cranking up the b/w > > > settings. > > > > You may want to raise the maximumThreads instead. Take a look at > > your localQueryTraffic diagnostic (at > > /servlet/nodestatus/diagnostics/localQueryTraffic/hour). If you're > > rejecting more than half the queries aimed at you, but you think > > you should be able to do more (not out of CPU or upstream bandwidth), > > then raise maximumThreads to some higher value. I was able to > > increase my node's effectiveness (at least in this metric) dramatically > > by raising it to 200. (It went from about 15% accepted to about > > 80%.) > > I am already running 1024 threads, and I have connect ration in the mid 80%. I > hardly ever see my node overloaded on the basis of anything other than > bandwidth, except that occasionally the routing time spikes up for a little > while. > > This is definitely a new fenomenon in the last 3-4 revisions in the unstable > branch. The bandwidth usage was as expected before - now it is about half of > what it was. > > Gordan
Maybe we need to adjust (or eliminate?) the fudge factor. There were some significant bugs in bwlimiting fixed recently, as well as the other issues. This calls for more testing. -- Matthew J Toseland - [EMAIL PROTECTED] Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so.
pgp00000.pgp
Description: PGP signature
