Well.. It happened again, now with version 539. # Current routingTime: 0ms. # Active pooled jobs: 300 (100.0%) # Current estimated load: 100.0%.
And my processor is doing nothing but idling... My node worked fine for a week or so, then last night I increased the number of threads from 200 to 300. Bandwidth usage increased by about 30% (which was also my goal with the change) and I thought that all was fine but today my node is doing absolutely nothing. A restart solved the problem this time also, bandwidth usage is now back to somewhere between 1 and 1.5MBit/s, but I don't really like the idea of having to do one or two restarts a day to keep my node working. /N -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] On Behalf Of Niklas Bergh Sent: den 25 november 2002 10:28 To: [EMAIL PROTECTED] Subject: RE: [freenet-support] Stale threads Going to be a little rude answering to my own post... Shutting down and restarting the node made it works as supposed again, connections are booming and processor and bandwidth is used again. Can it be so that the load estimation routine is sort of 'leaking' resources. Maybe it was a missguided effort from the overload protection mechanism that made my node unusable? Many of my threads where allocated to 'ConnectionJob', can it be so that the only thing they where doing was rejecting connections (also a kind of ConnectionJob I guess)? /N -----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]] On Behalf Of Niklas Bergh Sent: den 25 november 2002 10:07 To: [EMAIL PROTECTED] Subject: [freenet-support] Stale threads I think I'm seeing the (once before mentioned) problem with threads dying (or at least idling). My node has been running for 20 hours straight now. Requests have dropped from about 10k/hour to 100/hour, connections have dropped from 50-100 active to 3, incoming connections graph looks like an inverse logarithmic graph. My processor is unused (5% load) and there is no network activity worth mentioning. This is what the 'General information' page in fred says: Node Version 0.5 Protocol Version 1.46 Build Number 537 CVS Revision 1.90.2.13 Uptime 19 hours 9 minutes Load Current routingTime 0ms Active pooled jobs 208 (104.0%) [Rejecting incoming connections and requests!] Available threads 34 It's normal for the node to sometimes reject connections or requestsfor a limited period. If you're seeing rejections continuously the node is overloaded or something is wrong (i.e. a bug). Current estimated load 100.0% Note the extreme difference between my machine:s actual load (5%) and freds estimated load (100%) If I look under 'environment': Class Threads used freenet.Message: NodeAnnouncement 1 freenet.interfaces.PublicInterface$ConnectionShell 36 freenet.interfaces.LocalInterface$ConnectionShell 2 freenet.Message: DataRequest 5 freenet.OpenConnectionManager$ConnectionJob 188 freenet.client.InternalClient$ClientMessageVector 1 I think that 188 threads in ConnectionJob and 36 threads in ConnectionShell is quite a bit to high for my puny 3 active connections. Ideas? /N _______________________________________________ support mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/support _______________________________________________ support mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/support _______________________________________________ support mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/support
