I’ve been using Jellyfin for about a year. I’ve experienced many sudden issues with speed or connectivity, but they usually self-resolve over the course of a few days (I’d love to hear what that’s about).
Since the last major update, I’ve had intermittent speed issues. My network is a bit weird, but it’s what I have to go with for a while so bear with me…everything is wired cat7a as direct to the router as possible and broadcast exclusively through TailScale. My server (Win10) and another PC (Win11) are the combined shared storage, so I’m assuming one of the main points of failure between these 2 machines are to blame. In other words, the Win11 PC is acting as a shared network folder (where 2/3rd of my media is stored) AND a client (very inefficient I know, but it’s worked up until recently).
Today, I tried listening to a lossless song and it was taking about a minute to load 1 second of music. I’ve never had speeds that slow before.
The server’s hardware: -Intel i7-9700k @ 3.60GHz -RAM 16GB -NVIDIA GeForce GTX 980 Ti
Transcode settings enabled: -Hardware acceleration: Nvidia NVENC -H246 -MPEG2 -VC1 -VP8 -Enhanced NVDEC decoder -Hardware encoding
Task Manager for the server shows 95% (~10 Mbps) network usage to be to TailScale and Jellyfin. Network usage for the client is almost 0 Mbps. Memory usage for both machines is below 30%. CPU is less than 10% on both machines. Disk usage is even lower, 0% on both machines. The media in question is stored on HDD on the client-side machine (disk rated for >100MB/s read and write).
Without buying a new drive, NAS, or extra hardware, do you have any tips for troubleshooting my network to see if it’s something I can fix? Did I mess any settings up?
Don’t know if it’s relevant, but I have a Pi-hole and both machines have Simplewall.
I have no clue what could be causing that. I’d start looking into each link in the chain and making sure it’s working.
But any halfway reasonable config should be able to handle audio playback, no matter how lossless. Audio-only just doesn’t achieve datarates that would choke up… Anything.
Essentially, benchmark file transfers, transcoding, etc. Make sure each step of how it works is in fact working. Check drive SMART health… Whatever you can think of.
Also logs. No need to read through thousands of lines, but looking at the lines time stamped around when the issue occurs is always a good idea. FFMPEG logs, JF logs, client player logs, does SMB or whatever network drive protocol you’re using have logs? If it does, check em.
Outside of Jellyfin, the computers seem to communicate just fine. Transfer speeds are what I would expect and disk read/write is usually fine as well. Like I mentioned in the post, it self-resolved since I made the post. But my concern is the same: what could be causing the underlying issue? I’ll check the logs later and I’ll report back
Edit: forgot to mention that I have a few 5.1 SACDs @ 24/192k that have always completely choked my system. Doesn’t matter which machine or how extreme the transcoded settings are, it nukes the playback
So I checked the available logs after streaming a playlist for a while, which I’ve never done before (probably a huge mistake on my part). I have a generic “log_xxxxxxx.log” file that doesn’t seem to have any relevant info, but the only ffmpeg logs I have are related to video files from yesterday. Interestingly, the transcoding speeds seem to slow down over time; flac/wav files seem fine at first, but struggle to play after 30 minutes or so in a playlist of constant playback.
@LazerDickMcCheese @MentalEdge have you tools to monitor temperatures ? From what you’ve just said almost sounds like thermal throttling…
As short tests would give it a chance to cool off, while a heavy consistent load would slowly creep them temps up. Hwmon is what I used back in my windows days.
A friend had an nvidia card where the fan had stopped. He could play games fine for a short while (an hour or so) but after that it started to lag really.
I downloaded it and streamed audio that slows everything to a halt (5.1 SACD 24/192k). HWMonitor says the CPU is 30-34°C and GPU is 47-48°C with a hotspot of 60°C. My internal heat monitors show 30°C or less