I already have Jellyfin and Tdarr installed and configured. If I install the Sparkbox ARR stack, will it recognize my existing instance of Jellyfin and skip that portion, as well as use those config files to assist with the full stack integration? Hopefully that question made sense 😂
16 replies
tomspark wrote:
Good question, and it makes total sense! Short version: SparkBox's media stack installs its OWN Jellyfin and Tdarr as part of the bundle — it won't detect or import your existing ones or their config, so it won't 'skip' that part, it sets up fresh ones wired into the stack. The good news: the part that does the real work — the arr apps (Sonarr, Radarr, etc.) plus the downloader — doesn't care which Jellyfin you use. They download and organize everything into one media folder. So you've got two clean options: 1) Just use SparkBox's Jellyfin — simplest, it comes pre-pointed at the same folder the arr apps fill. 2) Keep your existing Jellyfin: install the SparkBox stack, don't use SparkBox's Jellyfin, and point YOUR Jellyfin at SparkBox's media folder. It'll scan it and your library shows up — your current config stays exactly as-is, it just gains the new folder. If you go with option 2, two things to watch: a port clash (both default to 8096) and making sure both point at the same media folder. Same story for Tdarr — SparkBox brings its own, use whichever you prefer. Want me to walk through option 2 once you've decided?
PinkyBroham wrote:
That makes sense. I’m on the fence, but I think the one that makes the most sense is option 1. Even though I did a lot of customizations to my Jellyfin, that’s nothing that can’t be duplicated. So I’ll just need to copy the movies I have already transcoded into the new media folders once they are created I’m assuming? I have a Ugreen DXP4800 Pro. I know you mentioned on one of your videos that you have tested some of those machines.
tomspark wrote:
Exactly right. Once you install the media stack, SparkBox creates the folders (Movies, TV, Music, etc.) inside its media folder — copy your existing movies into Movies and shows into TV, and SparkBox's Jellyfin will scan and show them. One bonus step worth doing: in Radarr, add that same folder as your root folder and use its 'Import Existing' option — then Radarr takes over managing those movies (renaming, quality upgrades) right alongside new downloads, instead of them just sitting there. Happy to walk you through that import when you reach it. DXP4800 Pro — great box, same UGOS platform as the ones I've tested, runs the stack well. One proactive heads-up since you're on UGOS: a few folks have hit a first-login permissions snag right after install (you'd see 'another claim in progress', or 'failed to save password'). It's harmless, and there's a quick one-line fix that makes it stick for good — if you run into it, just shout and I'll hand it straight over so you're not stuck.
PinkyBroham wrote:
Awesome I appreciate it. I’m gonna give it a go. This is the video I was going to follow to do the setup. Is there a Ugreen specific guide that I should be following, or is this video fine? https://youtu.be/Voeyw5mZ9Qo?is=9oEG7vE6ZVLifMKW
tomspark wrote:
Go for it! For the written reference, the always-current source is the guides at tomsparkbox.com/guides/ — those get updated as SparkBox changes, so I'd keep that open alongside whatever video you watch. The install itself is the same on UGREEN as anywhere: SparkBox detects it's a NAS and adjusts automatically, so there's no separate UGREEN-only path to worry about. The one UGOS-specific thing is just that first-login permissions snag I mentioned — and I'll fix you up on that instantly if it shows. If the video's a recent SparkBox walkthrough it'll be fine for the overall flow, but if a step ever looks different from what's on your screen, trust the written guide (and me) over the video — UIs move faster than videos do. Shout the moment anything looks off and I'll get you through it.
PinkyBroham wrote:
Good deal, will do. On a separate but related note, what are your views on the data privacy on the Ugreen machines? I know you addressed it in one of your recent videos a little, but I do see that data gets routed through Chinese servers. Was just curious on your opinion, as I know you are big on privacy.
tomspark wrote:
Fair concern, and you're right to ask. My take: the hardware itself is solid value — the privacy question is really about the STOCK software (UGOS), not the metal. Like most consumer NAS systems, UGOS phones home for its cloud/remote-access features and telemetry, and some of that infrastructure is overseas, which understandably gives privacy-minded folks pause. How I think about defending against it, easiest to most thorough: 1) The big one ships in SparkBox already: Pi-hole. It sits at the DNS level and lets you SEE and BLOCK what the NAS tries to call home to — you can literally watch it in the query log. That alone neuters most of the phone-home traffic. 2) With SparkBox running your stuff, your actual data — media, files, passwords — lives on YOUR box and never touches UGREEN's cloud. You can skip their cloud account features entirely. 3) The nuclear option, for zero UGREEN code: wipe UGOS and run plain Linux on it. I did exactly that on my own box — more involved, but then there's nothing of theirs left to phone home. So: good hardware, just don't trust the stock OS blindly — block it with Pi-hole, keep your data local with SparkBox, and go clean-OS if you want the belt-and-suspenders version.
PinkyBroham wrote:
Good info, thanks. I’m going to load up sparkbox and I’ll let ya know if and when I run into any snags.
PinkyBroham wrote:
Looks lIke I already ran into an issue. VPN Connection is saying unhealthy. I added the private key and the IP. Not really sure how to troubleshoot that at this point UPDATE: I figured it out. In the original config box, I saw a city very close to me, so I chose that, but it wasn't one of the actual servers for SurfShark. I typed in new York instead and it booted right up.
tomspark wrote:
Nice troubleshooting — that's exactly it. The VPN city has to be a real server location your provider actually offers; type a city they don't have a server in and gluetun can't connect, so it flips to 'unhealthy'. Picking a nearby major city (like you did with New York) is the move. If you ever want to pick the closest valid one, Surfshark lists all their server cities on their site under locations — but honestly 'nearest big city' works the vast majority of the time. Glad it booted right up — you're rolling now!
PinkyBroham wrote:
Is it normal for ugreen to hide the data/media folder from the native files app? I have to manually type in the path to the media folders each time.
tomspark wrote:
Yep, that's expected — not a bug. UGOS's Files app only lists folders it manages as official 'Shared Folders', and SparkBox creates its data/media folder as a regular folder, so UGOS hides it from that app even though it's there and working fine. Two ways to make it easier than typing the path: 1) Easiest: use SparkBox's own FileBrowser app — it points straight at your media folder, so you can browse, upload, and delete with no path typing at all. 2) If you specifically want it in the UGOS Files app (and mountable as a network drive on your computer): create a UGOS Shared Folder and point SparkBox's media at it. A bit more setup, but then it shows up everywhere natively — happy to walk you through it. And so you don't retype inside the apps: in Radarr/Sonarr/Jellyfin, once you add your media folder as a 'root folder'/library one time, it's saved — you pick it from a list after that, no typing. If those are making you retype each time, tell me and I'll sort it.
PinkyBroham wrote:
Finally back at it. Configuring the tdarr settings and when I run a test file for TV shows, it fails when trying to use the GPU to encode. With my old tdarr instance it ran just fine so I'm assuming it's a permission issue with the iGPU. The problem is with ugreen, I can only change the config in the project menu, however these are all set up in container. Any suggestions?
tomspark wrote:
Good news — SparkBox passes the iGPU into Tdarr for you automatically, but only when the hardware switch is flipped. That's almost certainly what's missing. Step 1 - turn it on: in SparkBox Settings set JELLYFINHWACCEL to intel (or over SSH, add JELLYFINHWACCEL=intel to /opt/sparkbox/.env), then: sudo sparkbox up tdarr That passes /dev/dri into the Tdarr container. Then in your Tdarr transcode plan pick a QSV or VAAPI (hardware) plugin and re-run the test. Step 2 - only if it STILL fails with a permission error (the device is there but Tdarr can't use it): that's the render group ID not matching your hardware. Find yours on the host: stat -c %g /dev/dri/renderD128 Then in .env add JELLYFINRENDERGID=<that number and run sudo sparkbox up tdarr again. Important: do this through SparkBox Settings / .env, NOT by editing the container in UGOS's own UI — UGOS changes get wiped on the next update, but the .env way survives. Give it a go and tell me where it lands (and which number stat prints, if you get to step 2).
PinkyBroham wrote:
I just used Tom AI and it kicks ass. Got me up and running and worked through a lot of roadblocks we ran into, mainly with my machine. Testing it out tonight
tomspark wrote:
Love hearing that — Tom AI's whole job is exactly this: walking you through the box-specific roadblocks that generic guides can't. Awesome that it earned its keep. Enjoy the testing tonight, and shout if anything else comes up!