[SSL Observatory] Tracker for data files

Chris Palmer snackypants at gmail.com
Sun Nov 29 11:54:38 PST 2015


As I recall, it was just the standard Python tracker, nothing magical. It's likely that the only problem is that web4 went away without an inventory of what services it was running. Perhaps all you need to do is run the standard tracker on a machine be and point the web4 CNAME at it.

https://noncombatant.org/

> On Nov 29, 2015, at 08:09, Martin Guy <martinwguy at gmail.com> wrote:
> 
>> On 29/11/2015, Peter Eckersley <pde at eff.org> wrote:
>> does anyone know if we can switch to magnet links that don't need a
>> tracker?
> 
> I think bittorrent clients should find others with the same files
> through DHT anyway.
> If I understand correctly, you should just be able to publish magnet
> links instead of torrent files, with the hash as the body of the URL.
> Or as well as.
> 
>>> I'm still (trying) to seed those files, but that obviously doesn't
>>> work to well if the tracker is down.
> 
> It might be worth putting one up at that address anyway, so that
> existing clients with the files will announce their presence, just to
> make things work better as they are. It shouldn't be that hard and I
> don't mind putting in the work to install and maintain it.
> 
>>> It seems that there are
>>> however still at least 5 people seeding it, there might be more if
>>> the tracker was actually still working.
> 
> Like I said, they should find each other through DHT and PEX but yes,
> startup would be faster for new clients. I like the fact that a
> torrent contains the file names, whereas with a magnet link it has to
> download that info; having them in the torrent allows users to select
> files to download before the torrent starts, otherwise they have to
> wait for the info to download when it finds the first other client and
> the user has to sit poised, ready to spring into action to the select
> the files at that point. Eck.
> 
>   M


More information about the Observatory mailing list