Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Failure reason "announcing too fast"] and rTorrent still downloads all files despite setting them off #930

Open
catzybluphish opened this issue Oct 24, 2019 · 8 comments

Comments

@catzybluphish
Copy link

When I torrent files with their manget link, say from nyaa.si, it'll initially show me the error [Failure reason "announcing too fast"]:
2019-10-24-193009_928x497_scrot
It'll keep showing that message until it starts after some time. I've compared downloading the same torrent with qBittorrent and there wasn't any issue on there. What does it mean and how can I fix it?

Also, I've tried to only download specific files I've selected from the files list. However, rTorrent will still create entries for those files, albeit with 0 bytes, as shown below:
2019-10-24-193232_1920x1080_scrot
The right window is ranger file manager showing entries for the entire torrent. How do I prevent rTorrent from downloading unwanted files?

I'm a new user so I, admittedly, haven't tinkered with the config much other than redirecting directory locations.

@leaty
Copy link

leaty commented Nov 15, 2019

I'm having the same issue, looking through my logs, whenever I add a torrent from a magnet link, it will announce and get a successful response. But then immediately after (1 sec) it announces again and gets [Failure reason "announcing too fast"], which then puts it into a terrible infinite loop of announce hammering every 2 seconds.

It didn't even get to use the data from the tracker - it's like it threw it out on the second announce, the only way to get it working again is by restarting rtorrent which for some reason makes it only announce once as it should.

@leaty
Copy link

leaty commented Nov 15, 2019

Here's a log showing what happens: https://pastebin.com/p4nv6fzs
Pay attention to the timestamps.

@kannibalox
Copy link
Contributor

what announce limits is the nyaa.* tracker subject to? rtorrent is only reporting back what the tracker tells it for that message. As for announce error handling, rtorrent follows a pretty well-defined pattern of backing off after each successive failure (up to 3 minutes, if i remember correctly).

@leaty
Copy link

leaty commented Nov 15, 2019

31 minutes, it lacks min interval - see output: {'complete': 953, 'downloaded': 21158, 'incomplete': 37, 'interval': 1889, 'peers': ...

I wish it did wait 3 minutes, but it's retrying every second, even after the first response as above. What ultimately happens is that I can see the amount of peers/seeds, but it refuses to connect to anyone while it keeps hammering the tracker. All up until it tries DHT after around 10 rapid announce requests, but it still keeps announcing. Or when I restart rtorrent, where it connects just fine to 40+ and also only announces once.

This sequence of events happens every time I add a new magnet link. I don't know what could be wrong, I used the following template https://github.com/rakshasa/rtorrent/wiki/CONFIG-Template and only changed some numbers in it.

@chros73
Copy link
Contributor

chros73 commented Nov 21, 2019

Have you seen this?

@leaty
Copy link

leaty commented Nov 25, 2019

@chros73 Yes, dht and most udp/tcp trackers do work just fine. But some do not, and when that happens - as with nyaa tracker, even if it receives a seemingly legit response (like in my previous comment), it tries again immediately after, where it receives "announcing too fast", after which rtorrent fires announce requests ruthlessly.

@pyroscope
Copy link
Contributor

When a tracker doing non-smart "too fast" responses meets rt's "repeat w/o exponential backoff" logic, that happens, yes.

@leaty
Copy link

leaty commented Nov 25, 2019

I haven't looked at the source yet, but if it's intended with a solid reason, then it's all good in my book. I assumed it might be a bug.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

5 participants