from Postmortal_Pop@lemmy.world to selfhosted@lemmy.world on 18 Jun 04:33
https://lemmy.world/post/48310412
If anyone else is having issues setting these up, message me and I will zip the whole build to you and walk you through setup.
UPDATE:
So I’ve made some major progress but still have a persistent issue. Radarr, sonarr, and lidarr are set to rename files, rename folders, and move them to the root directory. Even after importing the media the do not do that. I have both the boxes for renaming ticked, I have hard links turned off, they all have permissions for all the directories involved, and they have the media available in the program. If anyone knows how to fix this I’d love the help. I literally did all of this because I don’t want to manually rename 1600 files into a consistent scheme and Radarr apparently doesn’t want to either.
I used yams.media to do the full install. It was incredibly easy to use for most of the installation and setup. The Mullvad wireguard setup was a pain. The VPN part of yams specifically says to follow the instructions to the letter but the link it gives is a 404. The mullvad.md it was supposed to take me to was just “TLDR” and two code boxes with no explanation. I managed to bungle my way through with some knowledge from past attempts and the yams VPN test says I’m in Switzerland and my client is ready to go.
Yams wouldn’t let me set the directories I needed (it wants one directory for everything and I’m sorting them into different mounted drives) but it was actually remarkably easy to copy the yams config folders into my preferred directory and the yaml file directly into portainer to create a portainer stack running everything I needed. I even learned how to use the env and “advanced env input” in portainer to correct all the variable sections instead of writing all of them myself. All in all, it was exactly what I was wanting to do when I posted the TLDR.
Tl;dr: I understand docker is supposed to help get things running on different systems easily, can someone give me a copy of their working Arr stack?
Frustrated venting I’m past being new to this server thing having run mine for over a year so I guess I can officially say I’m just bad at it. I’ve been working on getting Sonarr, Radarr, and, lidarr running since 4 in the afternoon, discounting dinner that’s 6 hours of constantly failing to get these to work. This is my 5th time trying since I learned about it in April.
I’ve given up on the automatic downloads, I’ve given up on the request system, I’m even done with the torrenting, I’ll just do that on my phone. All I want is something that format my 5TB of media to Title (date) instead of MOVIE_TITLE_ALL_UNDERSCORE, or TB_1000, or movie.videoformat.year.special.deluxe.username.host.visit.my.site.please. I was sold on this idea that self hosting was a relatively easy thing that anyone can get into and while I have a good understanding of how a config.yml is supposed to look and work, and I’ve got a decent understanding of ssh and sftp between two computers, but trying to grt any one of these things to run is soul crushing. I literally work in the foster system and my worst cases do not give me the stress this does. I just want to get it fixed so I can watch Pokemon with my family and offer it to people who will never bother to log on.
Edit: OMFG I moved them back into individual folders and they work now. 6 hours of videos and tutorials and not a single thing saying they absolutely have to be in their own folders or it won’t work. edit unclear, brain stuck in toaster
Edit 2: turns out, Radarr can’t find movies at /movies/movie.mkv and needs /movies/folder/movie.mkv. Now Radarr can import movies but all other problems persist.
threaded - newest
Dont you want to look at something like
CasaOSZimaOS?It saves you the trouble of reading and setting up docker compose files
Glad you solved it yourself, but I’m still struggling to understand what happened, how did you have them all in a single folder if the filename for docker compose has to be one of a few predetermined things? I mean, you could have them all in a single file, which makes some things easier, but then you wouldn’t have been able to move them into individual folders. Would you mind explaining what happened there so that if someone else in the future has the same issue they might find the solution here?
Also, note that even if someone had given you an example of a working docker file you would still have to configure the service. For future reference, this site is great and has working examples of docker compose files for a lot of services, e.g. hub.docker.com/r/linuxserver/radarr
Finally, welcome to the club, sorry you had a bad experience the first time, it’s hard for us to know what’s obvious and what isn’t: xkcd.com/2501/
I’m very confused what it was that they moved into individual folders. And also configuring the naming of movies and shows is done in radarr and sonarr, not in docker compose.
I highly recommend Trash Guides for configuring these services. trash-guides.info
My only guess is that they were trying to map their config/downloads/library folders to the same location. And yeah, that probably wouldn’t work.
Poor explanation on my part. I had to move each of my movies from /movies/movie.mkv to /movies/folder/movie.mkv before Radarr could import them. I gave up after that so when I get off work today I’ll be able to actually import them and see if it moves and renames them this time.
Ah yeah, that’s because it will use the folder to store extra files as well, like subtitles, metadata, poster art, etc… Just dumping every media file into a single “Movies” library isn’t great, because you’ll have a ton of overlapping files. Separating them into their own folders allows you to store those extra files alongside the media.
I get that part, I just didn’t expect it to be incapable of starting with loose files. I’d like them properly foldered.
I see. I think you can do that, but it will need to be done during the initial setup. You could also try FileBot for the initial file renames, if you can’t get the initial file discovery and setup working.
Poor explanation on my part. I had to move each of my movies from /movies/movie.mkv to /movies/folder/movie.mkv before Radarr could import them. I gave up after that so when I get off work today I’ll be able to actually import them and see if it moves and renames them this time.
You shouldn’t be manually moving anything, though sometimes it is necessary. But when you’re first getting started I really just recommend following the TRaSH guides and then redownloading a few things to make sure it works properly. It explains a lot and it’s exactly what the people on the discord will tell you to do for all of this before going any further.
On the current install, absolutely. I had to manually move all my media off of my previous installation and onto external drives so I could fresh install the entire Ubuntu server OS because I wasn’t a fan of the logical volume system. I like being able to pull the whole hdd when sftp is too slow.
Sorry, it was late and I was very frustrated.
Radar, sonarr, and lidarr are each installed via docker into their own directories.
I recently did a complete reformat of my server and organization because the first try was a mess and this try I started organizing from the start. When backing up my media I moved just the video files from my first server to backup so I could delete all the extra jellyfin pics, torrent site ads, and folders of bulk subs. This meant the /movies was just 900GB of video files in a folder.
I mistakenly thought that Radarr would would take all the movies out of there and put them in the new location in their own folders as well as format the movie name and folder names to my specified schemes. Radarr would show the movies in the file list when I pick the directory but when I ran the import it would give me “all movies imported” without importing any of them.
As it turns out, each movie file needs to be in its own folder inside /movies, then Radarr will recognize they’re there and import them.
That’s all I managed to fix.
Ah ok this is the comment that needs to be in the main post. First things first:
You aren’t ‘installing’ radarr, sonarr, etc into any directories. They are containers, essentially entire operating systems located in a hidden folder on your server. You don’t ever touch these things directly, you only use docker commands to interface with them. There’s two ways to do that. Either directly running docker (
docker run linuxserver:radarr -p blah blah blah) or with a docker compose (docker compose up). The docker compose way is the ‘easy’ way to do it (actually the easiest is just using unraid and clicking install, but we’ll ignore that since so many people are telling you a billion ways to do things). Docker compose means you can specify all of your applications in a single file, and how they interact with each other. You will run one command to start all of them at once. And then they will read from whatever folders you configure in the service. This might be a bit confusing because up above you might see other people’s docker compose files and they specify things like this:sonarr: container_name: sonarr network_mode: "service:gluetun" image: ghcr.io/hotio/sonarr:latest volumes: - /mnt/drive/volumes/sonarr/config:/config - /mnt/nas/TV:/mnt/TV - /mnt/nas/Downloads:/downloads - /etc/localtime:/etc/localtime:ro depends_on: - gluetun restart: 'unless-stopped'and you would think that they’re configuring the sonarr locations for their tv and downloads, etc. But that is not what is happening. They are simply mapping a local path
/mnt/nas/TVto a path inside of the sonarr operating system/mnt/TV. This means that in Sonarr, in the web interface, you would configure the path/mnt/TV, NOT the path/mnt/nas/TV. You still have to configure EVERYTHING in the services themselves. All that docker is doing is setting up the operating system. Think of it like this, on a regular computer you can map network drives to letters like A, B, C, etc right? Well that’s exactly what you’re doing in docker, mapping ‘network’ (actually your main operating system) folders, to folders in the remote operating system (the one running in docker).In regards to having radarr rename things, you can have it do that, but you have to get the directory structure set up first, and you can run scripts to have nzbget or sabnzbd move things around for you. The experts on discord would be a much better help than most of us here I think, since they are all the devs on the project.
I actually do get all of this, I guess my issue here is not know the venacular to explain it?
I have the docker set up almost identical to that. I have an environment section with my user and group in it. My volumes are /home/user/docker/arr/radarr/config: /config /mnt/media1/1)unsorted: /downloads /mnt/media1/2)sorted: /movies. Inside radarr, add directory, /downloads shows the ~800 mkv files. When I go to import the movies in /download, it tells me there are no movies to import.
I fixed this by looping mkdir to put every file in /mnt/media1/1)unsorted into its own folder. Now Radarr sees all of them and imports them.
This hasn’t fixed literally every other facet of using them, I haven’t even set up sonarr yet. I’m still haven’t figured out gluten at all. Planning to run the yams thing someone else posted so I’ll see if that gets me anywhere.
What a wonderful rant!
The joy of struggle and learning! You learned 6 hours worth of what doesn’t work.
This is literally how I learn. Read, Do, Fuck It Up, ad nauseam until I get it right, and then write that shit down.
Not exactly your use case, but maybe someone will find it useful:
I really like compose.ajnart.dev - the creator of homarr (a dashboard with a lot of integrations) build a tool for creating docker compose files.
You can configure what services you’d like to run or use a predefined template (e.g. “the dad” with Jellyfin, Nextcloud, Pihole, Homarr and Home Assistant).
You’d still have to understand what you are doing, of course.
I’m down for something that makes the compose for me. I already get that part but it would be nice to not have to copy, paste, and fill my template each time.
Since it’s mentioned and this is a general suggestion, if you’re looking at Home Assistant, understand that the Docker version and the OS version are different and all of the videos don’t warn you about that. If you use the Docker version you have to manually install all the plugins where the OS has an app store. I didn’t know and all I wanted it for was chore-ops.
You should just install open claw + claude to manage and setup your services if you just want it to work and see a working example on your device. Its a simple task for AI and you can review and learn from configuration that applies to only your system.
Baring the obvious complaints about AI, my server is a dell tower from 2014, I do not have the resources to spare or the luxury of offloading my brainpower.
You can use a free account for limited usage.
absolutely do not do that. I can almost guarantee that any sort of AI will try to open a port that shouldn’t be open or in general expose you to a massive security vulnerability.
I made the mistake of trying to use deepseek to learn how to make a mod for vintage story. Suffice to say I’ll never make that mistake again.
Perhaps DockGE would help with your flow, it’s a docker web UI for managing services (in a basic manner, think a basic Portainer). It creates the separate folders for you, all you need to do is give the stack a name and plop the compose file in there. I use it to manage containers on all my VMs, mainly to have an easy overview in the browser but also to avoid issues like yours.
I’m actually running portainer at the moment and it’s been an absolute lifesaver in this nonsense. I’m not great in terminal and it helps me get the syntax right. Unfortunately, it doesn’t seem to do much to whip the arrs into working.
Haven’t tried it myself but can you run Nginx or Traefik in it, or would it give issues connecting to other containers?
As much trouble as any regular docker setup I reckon. It’s not doing anything special, just a nice front-end showing the logs, the compose and some details about the different services.
And many of my compose files in there make use of nginx, but the services nginx reference are all inside the same compose.
I am really glad to read stuff like this. Not because I like seeing someone struggle, but rather it makes me feel less alone that I am not the only one getting frustrated with things that seemingly work perfectly for everyone else that I run into every error/obstacle in the book lol. If I post for help in online forums like Lemmy, the replies make me feel so stupid and sometimes the replies are even condescending, which is extremely demotivational at times and makes me not want to ever ask to get it fixed. But eventually I get there and say screw all that and keep on going till it works. I’m happy you posted this and got your problem fixed my friend.
Also extremely relatable with the media server that friends and family refuse to use while paying like $30 a month on stupid subscriptions lmao
I get this feeling hard. You’re comment 3 out of 11 that I woke up to, comment one sent me to yet another guide to read and comment 2 congradulated me for getting it working. That’s on me for not classifying at midnight on a work night that the only thing I fixed was Radarr can now find the movies. Everything else is still broken.
Everyone talks about docker being an easy way to share things around, I’d assume it’d be easier to zip a working installation and send it my way than to find a guide I haven’t read.
Considering all of them are supposed to integrate with each other they’re relatively hard to integrate. I find it rather astounding they haven’t figured out service discovery.
We’re rowing the same boat
Hahaha so true. At least the struggling part. I spent a lot of time before getting my things running smoothly, but hei I was a full beginner. It 's not that I don’t dare too ask, it’s just that usually I try hard to find a solution before asking and mostly get it working. But yeah, that’s many hours of setup.
Last paragraph: definitely.
Late to the party but congrats on getting them up! Might I also suggest bazarr (github.com/morpheus65535/bazarr) so you don’t have to find subtitles manually.
I have that one too, it’s the only one that never gives me trouble. The rest are still broken as shit but atleast Radarr can find the movies now.
trash-guides.info/…/Docker/#docker-compose-exampl…
Trash guides are great and highly recommend by the servarr/sonarr team
That’s actually what I followed the previous try. It got me further than anything else but it fell apart because Mullvad doesn’t do OpenVPN and I couldn’t figure out how to get wireguard to work. Even then, on trying to rename and organize all my files, it just corrected them in Sonarr, Radarr, and jellyfin which broke all of my manually curated movies and it left all the files with the dumb mismatch they already have.
wiki.servarr.com/radarr/tips-and-tricks#creating-…
I recommend being really specific on what issues you’re facing when you want help. Your post was pretty generic and so you’re getting a lot of different answers.
If discord is your thing, the servarr/sonarr/trash folks each have their own discord where you can get help.
That’s actually what I was trying with the TLDR. As I understood it, docker is popular because once you get it running on one system you can use docker to make it run on any system. So my assumption was that someone who has the arr stuff working could just send me a copy of their config.yml and the directory it makes when you up and then I just tweak the locations and names on my side. I get that’s probably not possible but I’ll take literally anything at this point.
You’re not wrong but when you use somebody else’s config you use somebody else’s…configuration. Like if they use ProtonVPN, you’ll need to use ProtonVPN as well. If they use Usenet instead of torrents, that’s what you’ll get as well. If somebody uses Podman instead of Docker, etc etc. So this is why it can be more difficult than just ripping configs from strangers.
This is the classic problem where the more flexibility a program has, the more fragmentation comes out of it. The *arr stack is complicated for this reason. It’s a million different pieces that can be configured in a million different ways. Something like Nextcloud is much more plug-and-play. I’ve been doing self-hosting for years now and even I find *arr a chore to deal with.
Though nothing wrong with referencing other people’s configs to get a sense of what it’s supposed to look like. Start simple, look for somebody who has a radarr + qbittorrent + gluetun stack working, and go from there.
Yams.media is a script with an easy to follow guide that will hold your hand through the entire setup. All you need to do afterwards is import your existing media using radarr/sonarr.
Ok, so let me get this straight before I jump in. This is basically a prepackaged install wizard for docker, Arr stuff, torrent client, VPN, and jellyfin? I don’t need to install a different OS because it runs on Ubuntu server.
The one hang up I found is the same thing I didn’t understand using the trash guide, migrating gluten to wireguard for Mullvad. Could you or anyone that’s reading this parse this info for me and tell me how to double check if it worked?
github.com/qdm12/gluetun-wiki/blob/…/mullvad.md
There’s a guide for switching to Wireguard. Yams also includes a command:
that tells you if the VPN is working.
Not sure if that directly answers your question but there are a lot of step-by-step guides on the site and the author has also answered a lot of user questions on the message board.
Here is my docker-compose.yml file with sensitive info scrubbed, its been working for me for a few years now. It sounds like the problem you are having is not with Docker but something in your configuration once the container is running. Feel free to message me if you have questions.
services: gluetun: container_name: gluetun cap_add: - NET_ADMIN image: qmcgaw/gluetun:v3 devices: - /dev/net/tun:/dev/net/tun environment: - VPN_SERVICE_PROVIDER= - VPN_TYPE= - WIREGUARD_PRIVATE_KEY= - WIREGUARD_ADDRESSES= - SERVER_COUNTRIES= - DNS_ADDRESS= - HTTP_CONTROL_SERVER_ADDRESS= - HTTPPROXY_LISTENING_ADDRESS= - TZ=America/New_York ports: - 3129:3129/tcp # HTTP proxy - 8388:8388/tcp # Shadowsocks - 8388:8388/udp # Shadowsocks - 9047:9047 # Gluten http_control - 9046:9046 # qbittorent webui - 9696:9696 # Prowlarr - 7878:7878 # Radarr - 8989:8989 # Sonarr - 8686:8686 # Lidarr volumes: - /etc/localtime:/etc/localtime:ro restart: 'unless-stopped' qbittorrent: image: lscr.io/linuxserver/qbittorrent:5.1.4 container_name: qbittorrent network_mode: "service:gluetun" environment: - PUID=1000 - PGID=1000 - TZ=America/New_York - WEBUI_PORT=9046 volumes: - /mnt/drive/volumes/qbittorrent/data:/config - /mnt/nas/Downloads:/downloads depends_on: - gluetun prowlarr: container_name: prowlarr network_mode: "service:gluetun" image: ghcr.io/hotio/prowlarr:latest volumes: - /mnt/drive/volumes/prowlarr/config:/config - /etc/localtime:/etc/localtime:ro depends_on: - gluetun restart: 'unless-stopped' byparr: container_name: byparr image: ghcr.io/thephaseless/byparr:latest network_mode: "service:gluetun" init: true depends_on: - gluetun restart: 'unless-stopped' radarr: container_name: radarr network_mode: "service:gluetun" image: ghcr.io/hotio/radarr:latest volumes: - /mnt/drive/volumes/radarr/config:/config - /mnt/movies:/mnt/Movies - /mnt/nas/Downloads:/dCool I have different docker-compose.yml files for each service did not even think to put them in one.
I think they have to be for the gluetun(vpn container) dependency, but I could be mistaken. It does make it easier to
docker compose up -dand have the whole stack startup.I use wireguard so I wouldn’t know.
So do I, I’m just using it in gluetun so that I can pass all the traffic from this stack of containers in this yaml file through it and not other traffic. Gluetun is a pretty cool project, it made it easy too setup a connection to mullvad. github.com/passteque/gluetun
They dont. I have gluetun and qbittorrent in one docker compose and the starrs in a separate docker compose.
Here’s mine. I have separate stacks for media players (Plex, JF) and downloaders (sabnzbd, qbittorrent), so I added their networks to the config. I also chose to mount the volumes directly in the YAML instead of the VM’s
fstab, I found it plays a bit nicer that way. None of this is exposed to the internet. And I need to reconfigure the *seerrs, since Jellyseerr and Overseerr merged into one project…volumes: movies: driver_opts: type: nfs o: addr=192.168.1.175,nolock,soft,nfsvers=4 device: :/Movies tvshows: driver_opts: type: nfs o: addr=192.168.1.175,nolock,soft,nfsvers=4 device: :/TV_Shows music: driver_opts: type: nfs o: addr=192.168.1.175,nolock,soft,nfsvers=4 device: :/Music torrents: driver_opts: type: nfs o: addr=192.168.1.175,nolock,soft,nfsvers=4 device: :/Torrents prerolls: driver_opts: type: nfs o: addr=192.168.1.175,nolock,soft,nfsvers=4 device: :/Plex_prerolls books: driver_opts: type: nfs o: addr=192.168.1.175,nolock,soft,nfsvers=4 device: :/Books downloads: driver_opts: type: nfs o: addr=192.168.1.175,nolock,soft,nfsvers=4 device: :/Downloads services: sonarr: image: lscr.io/linuxserver/sonarr:latest container_name: sonarr restart: unless-stopped environment: - PUID=1000 - PGID=1000 - TZ=Etc/UTC volumes: - /var/lib/docker/volumes/sonarr_config:/config - tvshows:/TV_Shows - torrents:/Torrents - downloads:/Downloads ports: - 8989:8989 networks: - plex_default - downloaders_default radarr: image: lscr.io/linuxserver/radarr:latest container_name: radarr restart: unless-stopped environment: - PUID=1000 - PGID=1000 - TZ=Etc/UTC volumes: - /var/lib/docker/volumes/radarr_config:/config - movies:/Movies - torrents:/Torrents - downloads:/Downloads ports: - 7878:7878 networks: - plex_default - downloaders_default lidarr: image: lscr.io/linuxserver/lidarr:latest container_name: lidarrAcronyms, initialisms, abbreviations, contractions, and other phrases which expand to something larger, that I’ve seen in this thread:
3 acronyms in this thread; the most compressed thread commented on today has 9 acronyms.
[Thread #18 for this comm, first seen 18th Jun 2026, 15:20] [FAQ] [Full list] [Contact] [Source code]
I’m currently in my research phase, before I begin setting up my are stack. But I found this video, and it seems like it’s just the thing you’re looking for:
Ultimate automated media server 2026
In the description, he has this link to his GitHub for automated setup with all Docker containers needed etc https://github.com/loponai/arrstack
He also has a newer video on the area stack, haven’t watched it myself yet
I have watched this one, automation avenue is actually the one that got me the furthest before wireguard and gluten stumbled me. I haven’t checked the new one though.
Just use Saltbox
docs.saltbox.dev/saltbox/basics/basics/
It’s exactly what you’re trying to achieve but battle tested
What’s the difference between this and yam?
Haven’t used YAMS
But having read the page, I’d say it’s much more pluggable.
There are dozens of community plugins to install other docker based tools with Saltbox.
Also, it’s a rewrite of an older project and has been in development for longer than YAMS. I’ve been using it for around 8 years or so.
This may or may not help, but I went through the exact same sort of struggles that you did when I first started. Setting up all this was my first forray into Linux and learning about permissions and file systems was a hurdle but having conquered that my primary issues were always caused by one of the following:
Now that I actually seem to have a handle on all this…mind you, I’m no expert, but I can walk people through it in plain english. I eventually ended up switching to Proton VPN and use cloudflare tunnels to access my services from outside the network.
Yams (Yet another media server) starts you off with a full media arr stack, as well as your client of choice, torrent & nzb clients, and walks you through connecting all the pieces.
It also has built in methods for setting up a VPN tunnel for your torrents, which is definitely recommended. After running through the setup, you’ll end up with a compose file, and full docker media server stack. (I think even music arr client is included, but its been a long time since I set this up).
Does it also use VPN network for arr containers? So their torrent queries are also through VPN?
Not oc but just used it myself. It makes a gluten network in docker for the torrenting. It doesn’t run the others through it but you can modify the yaml to make the change yourself. If you want yams entirely it has a custom yaml that you can modify and a default that gets reset with every update. Modify the custom to change the network for the other arrs. Also, if I understand the system correctly, you only need to put prowler in the box with qbittorrrent. When using prowlarr, the arrs just say “hey I don’t have a thing” to prowlarr and prowlarr specifically finds the torrent and delivers it to qbit to download.
So I actually used it to set everything up last night and it did actually work all the way through, 9/10 easy. My only issue was that it expected a file structure that didn’t jive with the goblin tech I’m using for my server.
That said, it was surprisingly easy to modify the entire stack and copy it into portainer. I’m still working my way through the setup, but so far it’s smooth sailing.
just joining in on the misery. arr-s overpromise and underdeliver, tried for a month and gave up with just jellyfin and qbittorrent web. jellyfin is ok at masking the atrocious file names and directory structure, bless its soul
That’s what I’ll be doing if I can’t make it do the Renames. I just can’t handle seeing all that extra nonsense in the name files.