Warning: this might be one of my geekiest entries in a while. If you don’t care one whit about such things as internet protocols and development bugs, you’ll either want to ignore it, or read it and mock me.
After realizing that the two torrents I posted over the past 24 hours didn’t work at all, I started digging into the BitTorrent tracker I use (BlogTorrent) to see what the problem could be. After a lot of excavation, it turns out that BlogTorrent (and Broadcast Machine, its more mature cousin) is a little deaf to one of its configuration parameters, a parameter that’s likely to be less-often used, equally likely to be safely ignored in a bunch of circumstances, but is nonetheless both important and mandatory on network setups like mine. I posted a bug on the Sourceforce page for the tracker, so we’ll see what happens.
All this being said, I have to admit that I didn’t find it very easy to debug the entire BitTorrent conversation, a fact that made it much harder to find the problem and work out where it was located. BitTorrent itself uses a dead simple protocol, but I’m underwhelmed by the information about peers, seeding, and the like that’s available from BlogTorrent and the standard BitTorrent clients. I had to do a lot of webserver logwatching and launching of netstat dumps to figure out what was broken; I’d love to find a tracker which provides enough data so that I don’t need to repeat the performance.
I got the file no_times_picayune_20050830.pdf just fine yesterday.
• Posted by: James E. Robinson, III on Aug 31, 2005, 6:40 AM