Error sending Torrent to Downloader

[sumone78]

I recently added a new RSS feed/indexer to Sonarr and am getting the following error for releases it finds -

> Couldn’t add report to download queue. {release name}: Invalid torrent file specified

Looking through the logs, it doesn’t seem to mention what download link it is actually trying to use. However, here is the structure of the RSS feed, is there something that could cause this problem there?

<item>
	<title>{Some Show Title}</title>
	<description>TV/SD</description>
	<pubDate>Thu, 07 Jan 2016 04:43:33 +0000</pubDate>
	<category>TV/SD</category>
	<guid>{link to the torrent details/page}</guid>
	<link>{link to the actual torrent file}</link>
</item>

Also, not sure if it would make a difference but I noticed there is a strangely formatted portion of the .torrent file link. It looks like this -

https://www.{site-name}/download.php?id={someID}&SSL=1

With the strange text for the SSL var at the end there (&SSL=1). However if I manually copy and paste it into a browser it is able to download the .torrent file fine. I also tested manually adding that downloaded .torrent to uTorrent and it functioned fine.

Any clue what could be going wrong with Sonarr?

Edit: Some further detail from my logs:

MonoTorrent.TorrentException: Invalid torrent file specified ---> MonoTorrent.BEncoding.BEncodingException: Could not find what value to decode at MonoTorrent.BEncoding.BEncodedValue.Decode(RawReader reader) in m:\BuildAgent\work\6c3239faf2b92630\src\MonoTorrent\BEncoding\IBEncodedValue.cs:line 142 at MonoTorrent.BEncoding.BEncodedValue.Decode(Stream stream) in m:\BuildAgent\work\6c3239faf2b92630\src\MonoTorrent\BEncoding\IBEncodedValue.cs:line 99 at MonoTorrent.Torrent.Load(Stream stream, String path) in m:\BuildAgent\work\6c3239faf2b92630\src\MonoTorrent\Torrent.cs:line 701 --- End of inner exception stack trace ---

Edit 2: I used Jackett with the indexer to do a manual search, and was able to add the torrent to download client (via Sonarr) that way without a problem. So, there doesn’t appear to be any issue with the actual .torrent files from the tracker. Seems to be a bug with however Sonarr is trying to fetch the .torrent from the info it parses in the RSS feed.

[markus101]

Sounds like Jackett is fixing the URL as they handle that site specifically, but their RSS feed doesn’t work with Sonarr’s generic RSS indexer. This looks like an issue with the site’s RSS though.

[sumone78]

Well, I just double checked the Jackett class for that indexer. They actually don’t even use the RSS feed (a lot of index classes for Jackett don’t I’ve noticed). All their data comes from scraping the webpages/results, regardless of the source being ‘RSS’ or ‘search’ in Sonarr. Which, for the actual tracker website, the links are normal (no & in the link), so that would make sense that it is working.

Just a bummer since I’d prefer not to put the extra load on the site with making the actual page requests vs just RSS feed instead. Could some generic fix be applied to deal with the strange URL formatting, or is it too specific?

Edit: Just an update, the “weak link” of the process seems to actually be uTorrent. It blows up with the HTML encoded & in the link. That’s actually what is being passed up with the error in Sonarr.

Edit 2: Well… now I feel pretty silly. The download link itself actually can be parsed fine (&amp and all). The issue turned out to be the request needs to have a cookie as well, since the URL doesn’t include a passkey. This didn’t occur to me while testing because it was working fine in the browser (of course, only because I was signed in to the tracker there).

False alarm… sorry!