tomspark wrote:
Good news on the Intel side - you don't have to hand-edit anything or redo it each update. SparkBox already has a built-in way to hand Tdarr your Intel chip. Open /opt/sparkbox/.env, add the line JELLYFINHWACCEL=auto, save, then run sudo sparkbox restart media. That permanently gives Tdarr access to your Intel GPU, and because it lives in your config and not inside the container, it sticks through every future update on its own. After that you still flip the actual transcode settings inside Tdarr's own page (Hardware acceleration set to VAAPI, device /dev/dri/renderD128) - those save in Tdarr and persist too. On the CPU sitting at 100% even with everything paused: that's worth a look, not normal. Pausing the queue stops new transcodes, but Tdarr's library scanner and health-checks can still run in the background. Can you grab the tail of its log for me: docker logs sb-tdarr --tail 40 - that shows what it's actually chewing on. We're on 1.6.210 right now. Paste that back and I'll tell you exactly what's keeping it busy.