I noticed after every update, TDARR stops working properly and if I go into my libraries all of the plugins are errored out. If you try to search for new plugins they all show error's. If you restart TDARR to try and fix it then it wipes all of your settings and libraries.
9 replies
tomspark wrote:
That's frustrating, and it shouldn't happen — SparkBox saves Tdarr's settings, plugins and libraries to disk (they're mounted out of the container), so a normal restart shouldn't lose anything. Something's interfering, and a few quick outputs will tell us exactly what. Run these and paste them: cat /opt/sparkbox/state/overrides/tdarr.yml ls -la /opt/sparkbox/modules/tdarr/server sudo docker logs sb-tdarr --tail 30 The first is your Nvidia override — if it's got a stray volumes section it could be knocking out Tdarr's data mounts. The second shows whether your settings are actually saved on disk and owned correctly. The third shows what Tdarr is erroring on. With those three I'll know what's resetting it. (You're the same gdavidson26 who sorted the NVENC drivers, right? Good to see you back.)
gdavidson26 wrote:
Here is the output: gabriel@sparkbox-stack:~$ cat /opt/sparkbox/state/overrides/tdarr.yml ls -la /opt/sparkbox/modules/tdarr/server sudo docker logs sb-tdarr --tail 30 services: tdarr: runtime: nvidia environment: NVIDIAVISIBLEDEVICES=all NVIDIADRIVERCAPABILITIES=all total 20 drwxr-xr-x 3 gabriel gabriel 4096 Jun 6 01:41 . drwxrwxrwx 6 gabriel gabriel 4096 Jun 6 01:41 .. drwxrwxr-x 8 gabriel gabriel 4096 Jun 6 02:47 Tdarr -rw-r--r-- 1 root root 5 Jun 6 01:41 pgid -rw-r--r-- 1 root root 5 Jun 6 01:41 puid [2026-06-06T02:47:08.166] [INFO] TdarrServer - Node InternalNode [ID=7OMzYLlWT] [Address=127.0.0.1] [PID=280] registered [2026-06-06T02:47:08.166] [INFO] TdarrServer - Node registration handler: 15ms [2026-06-06T02:47:08.172] [INFO] TdarrNode - Node connected & registered, count:1 [2026-06-06T02:47:08.179] [INFO] TdarrNode - Server version: 2.47.01 [2026-06-06T02:47:08.394] [INFO] TdarrNode - Downloading plugins from server [2026-06-06T02:47:08.545] [INFO] TdarrServer - zipPluginsFolder took 1459ms [2026-06-06T02:47:09.643] [INFO] TdarrNode - Finished downloading plugins from server [2026-06-06T02:47:09.655] [INFO] TdarrNode - Running encoder tests [2026-06-06T02:47:09.692] [INFO] TdarrNode - FFmpeg version: 7.0.2-jellyfin [2026-06-06T02:47:10.028] [INFO] TdarrServer - Scanner test: Scanning file end [2026-06-06T02:47:10.028] [INFO] TdarrServer - Scanner test 1: FFprobe working [2026-06-06T02:47:10.028] [INFO] TdarrServer - Scanner test 2: Exiftool working [2026-06-06T02:47:10.028] [INFO] TdarrServer - Scanner test 3: Mediainfo working [2026-06-06T02:47:10.028] [INFO] TdarrServer - Scanner test 4: CCExtractor working [2026-06-06T02:47:10.029] [INFO] TdarrServer - ---------------Scanner tests end---------------- [2026-06-06T02:47:10.029] [INFO] TdarrServer - Running encoder tests [2026-06-06T02:47:10.051] [INFO] TdarrServer - FFmpeg version: 7.0.2-jellyfin [2026-06-06T02:47:10.380] [INFO] TdarrNode - Scanner test: Scanning file end [2026-06-06T02:47:10.380] [INFO] TdarrNode - Scanner test 1: FFprobe working [2026-06-06T02:47:10.380] [INFO] TdarrNode - Scanner test 2: Exiftool working [2026-06-06T02:47:10.380] [INFO] TdarrNode - Scanner test 3: Mediainfo working [2026-06-06T02:47:10.380] [INFO] TdarrNode - Scanner test 4: CCExtractor working [2026-06-06T02:47:10.380] [INFO] TdarrNode - ---------------Scanner tests end---------------- [2026-06-06T02:47:10.716] [INFO] TdarrNode - Downloading plugins from server [2026-06-06T02:47:11.001] [INFO] TdarrNode - Finished downloading plugins from server [2026-06-06T02:47:13.842] [INFO] TdarrNode - encoder-enabled-working,libx264-true-true,libx265-true-true,h264nvenc-true-true,hevcnvenc-true-true,av1nvenc-true-false,h264qsv-true-false,hevcqsv-true-false,av1qsv-true-false,h264vaapi-true-false,hevcvaapi-true-false,av1vaapi-true-false,h264amf-true-false,hevcamf-true-false,av1amf-true-false,h264rkmpp-false-false,hevcrkmpp-false-false,av1rkmpp-false-false,h264videotoolbox-false-false,hevcvideotoolbox-false-false,av1videotoolbox-false-false,libaom-av1-false-false,libsvtav1-true-true [2026-06-06T02:47:14.123] [INFO] TdarrServer - encoder-enabled-working,libx264-true-true,libx265-true-true,h264nvenc-true-true,hevcnvenc-true-true,av1nvenc-true-false,h264qsv-true-false,hevcqsv-true-false,av1qsv-true-false,h264vaapi-true-false,hevcvaapi-true-false,av1vaapi-true-false,h264amf-true-false,hevcamf-true-false,av1amf-true-false,h264rkmpp-false-false,hevcrkmpp-false-false,av1rkmpp-false-false,h264videotoolbox-false-false,hevcvideotoolbox-false-false,av1videotoolbox-false-false,libaom-av1-false-false,libsvtav1-true-true [2026-06-06T02:48:44.127] [INFO] TdarrServer - Job report history size is within limit. Limit:10240 MiB, Size:0 MiB [2026-06-06T02:48:44.128] [INFO] TdarrServer - Cleaning empty job report folders [2026-06-06T02:48:44.128] [INFO] TdarrServer - Finished cleaning empty job report folders gabriel@sparkbox-stack:~$
tomspark wrote:
Thanks — this is really helpful, and good news up front: Tdarr's actually running great right now. Your GPU is working (it passed the NVENC encoder tests), plugins downloaded fine, and your settings ARE being saved to disk in that folder — so a normal restart shouldn't lose anything. So let's pin down exactly when it wipes. Two quick things: right now, after that last restart, are your libraries and settings actually gone or still there? And when you say 'restart,' do you mean the restart button inside the Tdarr web page, or restarting it from the SparkBox side (or hitting the dashboard's Update)? That tells me whether it's the update step or something in the Tdarr app itself.
gdavidson26 wrote:
Whenever I update sparkbox and I go into TDARR, all my plugins say read error which then won't allow any transcoding or health checks to take place. I cannot add new plugins or try to fix the current ones. When I click the restart button for TDARR on the dashboard it then wipes all my settings and my libraries.
tomspark wrote:
Okay, that pins it down — thank you. Two things, and the first is the important one. Right now your Tdarr data is safe and sitting on your disk (I could see it in your earlier output), so before anything else let's make a backup copy so a future wipe can't actually hurt you: sudo cp -a /opt/sparkbox/modules/tdarr/server/Tdarr /opt/sparkbox/modules/tdarr/server/Tdarr.backup Second: until we ship a fix, please don't use the Tdarr restart button on the dashboard — that button is the exact thing clearing your libraries. If Tdarr acts up after an update, just leave it and message me instead. This is a real bug on our end, not you. I've flagged it to Tom Spark and we're digging into why an update leaves the plugins unreadable and why that restart wipes the data. I'll let you know the moment it's patched.
tomspark wrote:
Good news — we shipped the fix for this. Update SparkBox itself first with sudo sparkbox upgrade, then update your apps with sudo sparkbox update. After that your Tdarr plugins should stop going to read errors when you update, and the dashboard restart button won't wipe your libraries anymore. Your settings were always safe on disk, this just stops the update from breaking the link to them. Let me know it's behaving now and you can delete that Tdarr.backup copy once you're happy.
gdavidson26 wrote:
So I am still having the issue with it resetting my libraries in Tdarr but I noticed it might be due to the optional security updates for the apps. Sparkbox tries to do the optional updates every night but they always fail on Tdarr which rolls it back to the previous backup which removes my libraries. Not sure why it always fails on Tdarr during the update though. [0;34m[INFO][0m Updating module: tdarr failed to parse /app/state/overrides/tdarr.yml: yaml: line 4: mapping values are not allowed in this context Update command failed (exit 1) — failed to parse /app/state/overrides/tdarr.yml: yaml: line 4: mapping values are not allowed in this context. Attempting rollback... Rolling back to previous images... Restoring sb-tdarr... Rollback restart failed: failed to parse /app/state/overrides/tdarr.yml: yaml: line 4: mapping values are not allowed in this context
tomspark wrote:
Good news — we shipped a proper root-cause fix for this in v1.6.263. The wipe came down to a file-ownership mismatch: on a restart or update Tdarr couldn't read its own database, so it errored out and built a fresh empty one — taking your libraries with it. That also explains exactly what you spotted: the nightly app-update would fail on Tdarr and roll back into the wipe. Now SparkBox repairs those permissions automatically right before Tdarr ever restarts or updates, so it can read its database and the libraries stick. Run sudo sparkbox update to get it. If it ever wipes again after that, tell me and I'll jump straight back in — but this should finally put it to bed. Sorry it took a couple of rounds.
Chris wrote:
Taking you up on that, unprompted — because I think we missed something sitting in your own paste. Your state/overrides/tdarr.yml is invalid YAML: environment: is indented one level too deep under runtime: nvidia, and the two NVIDIA lines have no indent and no dash markers. That is exactly the 'line 4: mapping values are not allowed' error you hit, and it fails every Tdarr update, which is what triggers the rollback that clears your libraries. Corrected file: services: tdarr: runtime: nvidia environment: - NVIDIAVISIBLEDEVICES=all - NVIDIADRIVERCAPABILITIES=all Then sudo sparkbox up. Tom has since shipped the deeper upgrade-wipe fix too, in v1.6.408 — sudo sparkbox upgrade. — Chris