A single transcoded stream.
I’m sure it’s not a lack of Ram memory.
So, after some more poking around, it seems this issue was brought up with Intel in this Github issue. However, it looks like Plex is using it’s own self-compiled version of Intels compute runtime.
Another thread here on Linuxserver’s Github image of Plex has this exact issue/error.
While yes there have been issues with Intel’s drivers sometimes, could Plex simply be using an outdated version of Intel’s compute runtime here?
I thought that 21.x.x versions were needed too but it quickly became abundantly clear we needed to keep updating as Intel started releasing CometLake, TigerLake, RocketLake, AlderLake, and now RaptorLake (13xxx) processors.
Yes, Plex does compile the libraries. They’re compiled to be shared objects which the transcoder can load on-demand after it starts. Doing it this way eliminates the need to “install” new codecs every time there’s an update.
We keep in sync with the Intel releases.
If we were behind, there would be no RaptorLake (UHD7xx) support.
From our QA system
[chuck@lizum ~.2054]$ ssh plexqa@av1
Welcome to Ubuntu 22.10 (GNU/Linux 5.19.0-35-generic x86_64)
* Documentation: https://help.ubuntu.com
* Management: https://landscape.canonical.com
* Support: https://ubuntu.com/advantage
4 updates can be applied immediately.
To see these additional updates run: apt list --upgradable
Last login: Fri Mar 17 23:24:43 2023 from 10.2.15.100
plexqa@plexqa-av1:~$ uname -a
Linux plexqa-av1 5.19.0-35-generic #36-Ubuntu SMP PREEMPT_DYNAMIC Fri Feb 3 18:36:56 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
plexqa@plexqa-av1:~$ grep VERSION /etc/os-release
VERSION_ID="22.10"
VERSION="22.10 (Kinetic Kudu)"
VERSION_CODENAME=kinetic
plexqa@plexqa-av1:~$ cat /proc/cpuinfo | grep 'model name' | uniq
model name : 13th Gen Intel(R) Core(TM) i5-13400
plexqa@plexqa-av1:~$ date
Sat Mar 18 09:48:39 PM EDT 2023
plexqa@plexqa-av1:~$
Ah I see, after you mentioned the intel media driver I was looking through Intel’s updates to see if there had been mention of a similar issue or fix. From that previously linked comment it looked like Plex was using a 10+ month out of date version, but that seems to not be the case.
I wonder if there’s anything I could try that could determine where the problem lies (with Plex or the driver itself).
I just found a one of the known issues with transcoding.
Specifically –
-
With Nvidia 525.60.13 (more than new enough)
-
PMS 1.31.3.6819-2ef591a4c
-
Start PMS
-
Allow PMS to settle (inotify table to populate)
-
Start playback of 2160p HDR in Plex/web ( @ original quality )
– PMS probes Nvidia P2200 to determine if can transcode.
– Test passes
– PMS / Transcoder launch playback session flawlessly -
Took
date;nidia-smiconsole log @ server to backup PMS logs -
Wait 20 seconds
-
Change resolution to 1080p - 20 Mbps
-
Take another ‘date ; nividia-smi’ snapsnot
– caught transcoder exiting and PMS probing (as it should) in preparation for resolution change -
Plex/web shows no video
-
another ‘date; nvidia-smi’ snapshot shows no transcoder.
-
Stopped playback in Plex/web
-
Downloaded logs
Recreation of frequent complaint FINALLY captured.
Submitted to Engineering wtih Logs and Console session ‘date;nvidia-smi’ info.
We got this 
1 Like
Is the screen shot showing transcoder details from Tautulli or is that info in the Plex client somewhere? I don’t think I have seen that in my dashboard.
That’s the standard Plex/Web “Dashboard” activity indicator in the upper-right corner
Clicking it expands to show the Dashboard playback activity.
This symbol (to the right) allows you to expand or reduce the level of detail.
1 Like
WOW, I never noticed that thing! You learn something new every day.
@ChuckPa Any update on the Gemini Lake processor issue (J5005 and J4105). You had mentioned a while ago that the engineer found where it was having issues. Any progress? Thanks.
Yes, we worked through a different set of tests on Thursday.
We can see where the chip is reporting “Error -5” (Out of OpenCL resources)
At that point, I had to step back because I don’t know the code.
1 Like
Ok thank you! Appreciate your tireless help. Hope you’re feeling better btw. Looking forward to the next update.
I updated the server to the latest beta version (1.31.3.6868). Unfortunately still no change.
Unraid folks -
Need help
I think I broke something. Everything I read said to check the filesystem
Put it in another machine . fsck was OK.
Docker selection list – EMPTY
Is this easy to fix ?
Could you be more specific? Your docker containers have disappeared for some unknown reason? If so, are you able to reinstall the desired container via the apps list?
Thank you..
-
I think this was a “student driver” question and I turned the wrong way
-
I see the containers (apps) I am looking for under the “apps” tab
-
Question: Should “Template” (as I show in the screenshot) always be empty ?
(the list is always empty – except for the apps/containers already installed ) -
This is what I currently see but don’t understand.
( tried to screencap it but hotkeys and dropdown don’t work with each othere )
In addition to the default templates, you should see all the templates that were used to create the docker containers.
as you can see, I show <none> for [User templates]
@ChuckPa any updates on this matter?
There is no progress report from the engineer who was assigned.
I will ask tomorrow in our team meeting.
I have the answer.
- To resolve this requires the Intel Media Driver to be updated. (trivial)
- Updating the Intel Media Driver for FreeBSD is an extremely painful task.
Waiting on the management decision to change product offering for FreeBSD.
1 Like
