Server Version#: 1.18.0.1846
Player Version#: 4.10.1
Every now and then when I’m streaming a video (loads quick, no buffering, no stuttering) and decide to seek, it will sometimes seek perfectly fine regardless of where you jump to in the video, but sometimes it will get in a perpetual buffering loop that does not stop until you stop the video and restart it. The following repeats in the log (+ some other identification such as user, device, profle, etc.) over and over. It does not seem to matter if you stream it locally or remotely, on a TV, FireTV, Chromecast, Desktop… every platform seems to do it.
Oct 22, 2019 22:21:12.895 [0x812f67200] DEBUG - Request: [x.x.x.x:63702 (Subnet)] GET /video/:/transcode/universal/dash/x5hva2fri8qxjfdpinkp3c42/1/51.m4s (8 live) TLS GZIP Signed-in
Oct 22, 2019 22:21:12.895 [0x812f67200] DEBUG - Asked for segment 51 from session.
Oct 22, 2019 22:21:12.895 [0x812f67200] DEBUG - Returning segment 51 from session
Oct 22, 2019 22:21:12.895 [0x812f67200] ERROR - Couldn't find the file to stream: /usr/local/plexdata-plexpass/Plex Media Server/Cache/Transcode/Sessions/plex-transcode-x5hva2fri8qxjfdpinkp3c42-ac17a54f-3a72-4642-a3cf-765aeb7dd8d3/init-stream1.m4s,/usr/local/plexdata-plexpass/Plex Media Server/Cache/Transcode/Sessions/plex-transcode-x5hva2fri8qxjfdpinkp3c42-ac17a54f-3a72-4642-a3cf-765aeb7dd8d3/chunk-stream1-00052.m4s
Oct 22, 2019 22:21:12.896 [0x80b540900] DEBUG - Completed: [x.x.x.x:63702] 404 GET /video/:/transcode/universal/dash/x5hva2fri8qxjfdpinkp3c42/1/51.m4s (8 live) TLS GZIP 0ms 410 bytes (pipelined: 251)
Have you managed to fix this or narrow down the cause?
I’ve had this issue on Ubuntu and now on Arch for as long as I’ve used Plex, it’s starting to become quite a hassle having a restart a video 5 times if you want to seek back (or foward).
This is what I see in my log:[0x7f12de7fc700] ERROR - Couldn't find the file to stream: /tmp/Transcode/Sessions/plex-transcode-d57k64ui8ud3angf0sk3bx0z-93b218b9-1d58-48d5-820d-4e794bfd4b93/init-stream1.m4s,/tmp/Transcode/Sessions/plex-transcode-d57k64ui8ud3angf0sk3bx0z-93b218b9-1d58-48d5-820d-4e794bfd4b93/chunk-stream1-00157.m4s
It appears as if the file really is nowhere to be found. And this is if I return back to a time in which the video was playing correctly:
[0x7f1369443700] WARN - JobManager: Could not find job for handle 60483
It is at this point where the stream must be restarted or there is never a change in the video starting or stopping, no matter where I seek to on the timeline.
The tag on this issue could probably be changed from just FreeBSD to server-linux
I had this set originally to server-linux but a mod changed it to server-freebsd. Not sure why ¯\_(ツ)_/¯. It seems like it happens more often when streaming to a slower connection (such as over the web).
This is still happening on client 4.14.2 / server 1.18.2.2058
I can make this issue occour on every single video I play unfortunately. The server is local but the media is over an NFS share. This is never an issue for anything else I use the share for - and it always maxes out the 1Gbit bandwidth available.
Unlikely that the share is the cause, all my files are on the same server on a ZFS RAIDZ2 array. Server is running an i7-6800k, 32GB, no GPU. Might add a GPU in the future for better 4K streaming if I get into it. Happens on 480p, 720p, 1080p, and 4K files. Haven’t had it happen on audio files yet (mp3/flac).
For local streams it can happen on wired (desktop/web clients) or wireless connections (chromecast, roku, phone). For remote streams internet shouldn’t be a factor as I’m on fiber.
If it doesn’t do it immediately I can kinda “force” it to happen by jumping far along in the video then jumping back towards the beginning to a section that was not buffered previously. It’s definitely more difficult to do on a local wired client than a remote client on a slower (~50mbps) internet connection.
Edit: My ZFS array gets about 260mb/s read speeds.
Has anyone had any luck with this. I created a topic with the exact same symptom but with a completely different HW/Software combo and haven’t had any luck fixing it.
To add to my previous post in my troubleshooting I was able to confirm that according to windows the chunk fule that’s “missing” was never even attempted to be created so ther’s gotta be a flaw in the logic writting the file or keeping track of which file number to look for.
