In message <20021104124718.GA652 at sporty.spiceworld>, Oskar Sandberg <oskar at freenetproject.org> writes > >My node is doing over ten thousands qph now without breaking a sweat, so >I'm fairly certain the problem is indeed fixed (though this is >partially due to the fact that almost none are successful). 609 is a >little better still since I reversed a conditional in 608 so it was >still accepting bad announcements. > >"maximumThreads" is the correct name for the setting. Please do not use >the 0.6 as the overload threshold - with 80 threads that leaves only 12 >for the node to process requests and new connections, which is far too >little. The default value of 0.85 and 0.9 are good - changing those was >a typical example of going after the symptoms rather than the problem.
Can I just ask for an explanation of this? Build 608 (and all since 525), maximum threads -120 (or 120 or -60 or 60) my node is currently showing pooled threads = 250, and on previous form this will increase until it is not permitted by the OS to make more. Is this supposed to happen? > >I'm fairly confident that the new code is working as it should, and >hopefully Matthew can do a new release with it sooner rather than later. > >On Mon, Nov 04, 2002 at 07:20:30AM -0500, Ed Tomlinson wrote: ><> >> From my end it looks good too. Version 608 is the _first_ version since >> 525 that does not peg the cpu at 100%. It still seems to use more than >> 525 but not too the extent that it stops doing useful work - the node is >> processing about 800 requests per hour (about 8400 jobs). I am using the >> following in freenet.conf: >> >> maximumThreads=-80 >> maxThreads=-80 >> doLoadBalance=yes >> overloadHigh=0.65 >> overloadLow=0.60 >> >> I have seen both maximumThreads and maxThreads suggested on the list. >> Which is correct? >> >> TIA >> Ed Tomlinson >> >> >> > -- Roger Hayter _______________________________________________ devl mailing list devl at freenetproject.org http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl
