I am sure I am just missing something simple… I have prowlarr -> sonarr/radarr -> qbittorrent -> jellyfin I created three directories. /jelly/video /sonarr /radarr. I configured sonarr and radar to use their respective directories. And I configured qbittorrent to use /jelly/video as the default download dir.
But what seems to be happening is that if I download a movie, it ends up in both /radarr and /jelly/video. And then if I delete it from /jelly/video it doesn’t seed for others.
What am I missing here?
Radarr has the option to link the files when it imports them. Look in the settings
Use Hard links instead of Copy - Use Hard links when trying to copy files from torrents that are still being seeded
https://wiki.servarr.com/radarr/settings
https://trash-guides.info/Hardlinks/Hardlinks-and-Instant-Moves/
I’ve never managed to get it working. But I always have to do thing multiple times until it works.
Hardlinking works great for me 🤷♂️ but my Arr setup is on a paid hosting service so i imagine the virtual server was setup for that intently.
deleted by creator
The way I organized my setup was using a file structure like this:
- Videos
- Movies
- TV
My media player and torrent client have access to the videos directory, and Radarr and Sonarr have access to their respective directories. The *arrs add the files to the torrent client with the destination being their respective directories, and upon completion it triggers a media player library re-index. This way you can seed and stream concurrently.
So do you set the torrent client default download dir to videos? And the system is smart enough not to make a copy there because there is already one visible to the client?
Mine is a little more complicated, but it gives me piece of mind and the ability to see what each program is doing, and to manually sort files if sonarr/radarr stop working for whatever reason
My folder structure is
- downloads
- incomplete
- complete
- tv
- movies
- video
- tv
- movies
Each component of my stack is isolated using docker and can only acess what it needs to. Sonarr, Radarr and qbittorrent are configured to use labels to keep the downloads directory sorted.
I can post my docker-compose.yml file if you want to have a look.
deleted by creator
So you are intentionally keeping two copies of things?
I am keeping 1 copy, with a hardlink to the other. It gets removed from qBittorrent once it has finished seeding
oh, hardlink. Linux I am guessing then? I am on windows for now. And it has been years since I tried to make a link in windows. I don’t recall it going well back then. :) So what do you mean by finished seeding? Someone else implied they only seeded to some limit. What is the story there?
Yes, however Windows offers hardlinks too, you just can’t span them across drives with either os
- downloads
- Videos
That’s normal, Radarr makes a copy of torrent downloads into wherever for Plex/Jelly so that the download folder-version can seed. I can’t remember where the setting is, pretty sure it’s in Radarr but there’s a setting that specifies how long to seed and when it reaches the seed threshold to delete the file in the download folder.
So this implies there is something else I am missing. I assumed people would just keep the whole library open for seeding. Why would you want to delete the file in the download folder?
Many people do not want to seed downloaded content forever for storage reasons. In these cases, you would download the file with your download client and leave it in that download directory to allow seeding. It’ll be hardlinked to the Radarr/Sonarr folder for indexing, which does not use up extra storage space. Once a certain seeding/time goal is reached on the torrent, the torrent file will be deleted to make room for new torrents. This does (to my knowledge) not delete the file from the disk, meaning it is still accessible for your media center.
Especially for people who run their software on hosted solutions with limited storage space, this is important to do. If you have all your software running on a local server with (virtually) infinite storage, this is not as much of a worry to you. It is probably still in your best interest to use hardlinks instead of copies, to save on storage space.
You would delete the download to completey stop seeding. Take a look at the guides others have posted.
My structure is this:
torrent -movies -TV
media -movies -TV
qbit downloads to the torrent folder where it seeds out from. Radarr/Sonarr make a Hardlink copy from the torrent folder to the media folder. Your media software watches the media folder.
but why would you want to stop seeding?
I mean that’s up to you, my personal settings are 2:1, it’ll delete and stop seeding once it’s uploaded as much as 2 downloads worth.
But it’s important to remember that Radarr/Sonarr are Usenet first, torrents didn’t have official support till fairly recently and it’s still a little wonky
Officially, to seed you basically have to have a duplicate file until your set threshold is reached.
If you want to seed forever (or at least a really long time (ty btw for your contributions)) I’d say you’re probably going to want some custom scripts. Have Radarr move the file and rename as normal and then your script to symlink it back to the torrent directory under the original filename so it can continue to seed without taking up double space for every movie
torrents didn’t have official support till fairly recently and it’s still a little wonky
I don’t think this is true at all. They’ve both had solid torrent support for years, across multiple major version numbers. It’s neither wonky nor recent.
I’d say you’re probably going to want some custom scripts. Have Radarr move the file and rename as normal and then your script to symlink it back to the torrent directory under the original filename so it can continue to seed without taking up double space for every movie
Further driving home that this dude is full of shit, hardlinking the files is enabled by default in both Sonarr and Radarr and certainly doesn’t require any custom scripting.
OP, quit listening to random people online and spend some time reading the documentation yourself.
