For the short time it was up, torrents.textfiles.com was pretty popular. The idea behind it was good, and still holds: every once in a while, people get really excited and try to download all of textfiles.com. The site is many gigabytes (assuming you go for some of the sub-sites as well) and so this takes a while. Unfortunately, the massive number of small files and the fact that some people use some pretty poorly-working clients to attempt massive parallel transfers of the site means that my bandwidth gets nailed.

Here’s what nailed bandwidth looks like. The blue line is the outgoing, and the green line is the incoming. The reason the blue line is sometimes flat at the top is because they have maxed out the T-1 that TEXTFILES.COM and the related sites run on.

There’s really no need for this to be happening. Sometimes, people are pulling things down and they find the throughput to be less than enjoyable, but that’s life when something’s free; that’s why I’m going to be becoming a non-profit organization this year.

For people who want “THE WHOLE SHEBANG”, a bittorrent connection is the obvious solution. Then everyone who wants multi-gigabyte hits (actually, the main TEXTFILES.COM archive works out to about 700mb) can just grab it that way and help each other along the way.

The thing is… bittorrent seems to have some major issues with the way I want to use it. So I’d like some help.

The way I want to use it, Bittorrent would replace an HTTP link or an FTP link. You would go to torrents.textfiles.com and there would be a set of files offered, where you would grab the torrents and then you download them. I would be the seeder, I would be offering up the new seeds, and you’d choose from them.

Basically, I want torrents.textfiles.com to be almost exactly like LEGALTORRENTS.COM.

The thing about a lot of tracker/bittorrent sites is that they are very simple to use from the user-side. If you have the client installed (and really, it’s on the level of Flash to install a bittorrent client) then you click on the magic torrent link, it asks where you want to dump the file or directory, and “off it goes”. There’s a number of clients of varying usefulness, but they all kick in nicely. The issues are on the server side.

Part of it is the webpage. I like how it shows you statistics of how much has been transferred and how many people are using it. I understand this tends to require PHP, because that’s how many people think now, and that’s OK. I do not think MySQL is OK.

I understand that I have to run a tracker, but I do not understand why trackers are so difficult to deal with in UNIX; they all want to shoot stuff to the screen and I can’t find a way to do an easy tracker startup script. It’s all hopelessly complicated for the relatively easy thing I want done (start a tracker, serving these files).

If someone could help me, I would appreciate it, and textfiles.com would be a better site for it.

Update: A number of folks have assisted me, and my own research was redirected in several helpful ways.

Ultimately, the solution I am going with is to use the original python bittorrent client, with a script that redirects the logging output to disk, and a startup script which does all this setup on reboot. For the nice torrents.textfiles.com webpage, I will be using a nice simple PHP script called Torrent Trader Lite, which does all I need it to do, without requiring MySQL or a bunch of other stuff. I will be creating torrents shortly. Thanks again.

Categorised as: Uncategorized

Comments are disabled on this post