Which kernel are you running ? System?
Ubuntu 20.04 server here
Could I add the kernel :
Sudo dpkg -i *.deb and then im good to go? 
I wouldn’t 
.
I’d just add the repo and apt install it.
If you still have issues it might be because of some intel/mesa package I have installed as well (but I would start with just the kernel)
A HUGE THANKS TO @christophedc0
Seems to got it working now
Kernel updated to 5.14.12-051412-generic
Installed the Intels updates from post 307 .
At the moment running 5 transcodes: 4 movies of 1080p converted to low quality so the server will transcode and 4K HDR movie transcoded to 1080p . HDR Tonemapping disabled.
So far running SMOOOOTTTHHH - NO vacuum cleaning from my NUC
CPU load average: 0.19
1 Like
That’s nice to hear.
did you have to use the intel modules or was the kernel update sufficient ?
Ps: I got HDR tone mapping enabled and still haven’t seen an issue with it.
I added them all , just to make sure . Since you had success with your setup at post 307, i copied it and tried it out.
For now , so far so good. So l hope that it did the trick
Thanks for all your help in pm @christophedc0
I just tested on Ubuntu Server 20.04.3 LTS (GNU/Linux 5.11.0-37-generic x86_64) with a i5 11600k and no more green artifacts!! So looking forward they post it to the beta channel now!
1 Like
I had a server crash last night - enabled logging today just in case if it happends again.
Had 3 hw transcoding on going last night, I was watching a movie and all of a sudden , Plex couldnt read the file of some sort. Then the server went offline and hang in systemctl.
No vacuuming so far though but the server was a bit hot when I logged in.
Which logs should be enabled?
I have “Enable Plex Media Server debug logging” but not verbose logging. OK?
Hmm transcoder is dying.
A user of mine reported that the stream stops - and then I can see hw transcoding is gone.
Vacuum cleaning is back.
Had to roll back to the official release.
I have too many problems with the beta 
Any GPU crash ? Or just related to the experiment build ?
I have to guess that now the remaining issues are driver related and we have to wait for Intel to fix them
Issues · intel/media-driver (github.com)
I only had 1 crash happening, not sure why, but it was when someone logged in from distance.
vcs0 reset error, had to reboot the whole server.
since then I disabled HDR tonemapping
No crash since ?
Well I don’t know to be honest.
The server runs smooth - everything runs perfectly . I even tried 6 transcodes at home today, and the server is running fine for like 20-25 minuts… maybe more or less. ( Was working at the same time )
After a while I have the spinning buffering thing, and the server needs to be restarted.
Earlier today I testet with a friend. He was testing if remotely. I saw (hw) transcoding was running, but after a while I also heard the server spinning up the fan rapidly like it was getting really hot.
I logged in SSH and saw CPU temp was about 100c .
The (hw) disappeared and he was transcoding directly on the CPU . I saw the CPU in htop was increasing to 80-90ish%.
Mine is hitting 100 degrees very often too, it’s common to these “laptop” APUs to run hot, as their radiator and fan are tiny.
I suppose PMS is converting the buffer length as quick as possible and could hit the CPU/GPU heavily with both video+audio transcoding, and then it will “throttle” the encoding to the video playing speed, causing the load to flatten over time.
Up till now, no issues. (±36hours later)
I have data reporting running on the nuc, that monitors the igpu, cpu, temperature and docker memory.
It runs average at 45°c and max has been at around 75°C (I’m running in balanced mode)
Dumb question, but this is only available in versions 1.24.6+, right? I tried docker version 1.24.4-5081 (which is the same hash as latest), and was unable to get this to work on the 5.14 kernel with Intel packages installed on the host. My build is a bit different… I’m exposing this through kubernetes and my hosts have the i915 flag available (works on older generation NUC). Plex runs fine on IrisXe hosts where it didn’t in previous builds (would crash continually), but I’m not seeing hardware encoding kicking in at all with the options enabled.
Edit: Why are plex changelogs so hard to find?
36 hours with no crash is very good, I never got the chance to have a 36 hours uptime because of the issue but it looks like it’s fixed to me (Finally I would be able to watch HDR content on some SDR screens, yay)
The server runs smooth - everything runs perfectly . I even tried 6 transcodes at home today, and the server is running fine for like 20-25 minuts… maybe more or less. ( Was working at the same time )
It might be something on your side, when the crash happen in my side, Plex just tells me that Plex cannot convert for whatever reason and video player closes, direct play is still possible and no wheel of death report
I do have some buffering wheel of death issue but it’s unrelated to this, maybe related to hard drive enterring sleep mode when it should not ?
but this is only available in versions 1.24.6+, right?
It’s something apart from public and beta as far as I know, it’s just a “Forum preview” currently
I have now updated with a new intel driver , thanks for info @christophedc0
A buddy testet external streaming, and he got an error.
Transcoder failed , conversion failed and transcoder exit due an error.
But which log file do I have to search for info , and what do I have to look for?
Im currently testing a file on my own via Chrome , so far so good, no crash or errors.
Installed intel-media-va-driver instead of none free version.
I’m able to replicate a crash all the time.
Reported it in the preview topic:
2 Likes
