Registered Member
|
After installing 2.1beta, I noticed that my 150KB connection was getting saturated by KTorrent, even though KTorrent was showing only 20KB/sec download, in accordance with the bandwidth limit. After much investigating, including installing pktstat (in Debian), I discovered that HTTP connections to prq.to trackers that were HTTP GET'ing "/scrape?infohash=..." were going on forever, downloading megs and megs of data, and usually at least three of them were going at the same time. I stopped all torrents and started the ones that were not on prq.to trackers, and the problem went away. After starting just one torrent on http://tpb.tracker.prq.to/announce, this connection was started which saturated my connection until it had downloaded about 7 megs:
(rate) (% of connection) (total data) (protocol) (local address) (remote address) 133.4k 66% 7.2M tcp 192.168.0.2:47433 <-> 85.17.40.34:80 └ 200 GET /scrape?infohash=%1f%f8%8c%a0%03%93%8f%1d%d1%fb%95%01%ab%19%afq%bbE%07%98 Then I started another torrent on that tracker, and another connection just like that one was started, and again it downloaded megs and megs of data. Both of these torrents had zero peers and zero seeds at the time. It looks like what KTorrent is doing is, since the torrents have zero peers, it's rescraping the torrent every few minutes. But for some reason, the scrape results in downloading 7+MB of data. With four torrents on that tracker, each with zero peers/seeds, each with 7+MB scrapes, and each being rescraped every couple minutes, KTorrent continues to saturate my connection with non-torrent data. I've no idea why the scrapes are so big. I don't even know if this is definitely a KTorrent bug. But something's wrong here. |
Moderator
|
|
Registered Member
|
|
Moderator
|
I have tons of torrents lying around on my disk, the only problem is to find one with this tracker. Allthough, I can probably just add the tracker to an existing torrent. |
Moderator
|
|
Moderator
|
Registered users: Bing [Bot], Google [Bot], q.ignora, watchstar