Please try 1.29.2.6364 AND 525 Nvidia.
Others and I’ve been doing testing. We might have it isolated.
being unrealistic!! wauw…I work in a IT company, and if we did this to “PAYING” clients, they would kick us out…this testing should have be done in pre-prod before any release has been made!..
I will try, and sorry @ChuckPa it’s not an attack on you…just strange that PLEX are not more official about this technical issue.
We have spend 5 days trying to fix this server issue,…
I’m working with our new management to change that.
I’ll share now, in advance, of what I personally am contributing to the project.
We’re building a fully automated, full-sweep, test of transcoding. (I’m doing all the scripting work).
Should it have existed long ago? Yes. I can’t speak to why it doesn’t already exist.
Will it exist and be used when I’m done? Yes.
Since I’ll have access to it, I’ll be able to take everyone’s submitted sample(s) and recreate them on a known test bed system.
Additionally, QA will also be able to run their tests as well.
My personal goal is 100% regression and new transcoding testing of every release before it gets put up in PMS Beta status.
Might this develop into more? It would be nice but small steps at a time.
(In my former life, I had to run my builds for 7 days using pre-recorded and 7 days of live satellite data then verified before the build was signed off and submitted to QA)
6 Likes
Hi there,
@enjoyiptv I work in IT too but there is not a lot of software out there that you will pay with a lifetime price, no? 
@ChuckPa no I’m not dead, I just had some struggle to test those last few days… I’ll still try 
Thanks. I know real life takes priority. Whenever you get a chance is greatly appreciated.
1.29.2.6364 and current Nvidia drivers should be just fine.
can confirm that PLEX 1.29.2.6364 with Nvidia 470 drivers on ubuntu 20.04 is doing GPU transcoding on all conversions 
Downgraded nv to 515.86.01 on pms 1.31.0.6654
errors persist
CUDA_ERROR_INVALID_VALUE: invalid argument and
progress/log?level=0&message=%5Bhevc%20%40%200x7f7e778389c0%5D%20No%20decoder%20surfaces%20left (14 live)
Running on docker so using the release candidate deb or downgrading pms is a big stretch. Limping along with CPU transcode waiting for a rollback on plexpass docker image while you guys debork this release or fix to come out.
Hello!
I downgraded my Plex version to “plexinc/pms-docker:1.29.2.6364-6d72b0cf6” and let the database migration run for 3-5 minutes and was able to watch content that wasn’t loading on the latest version.
I’m using the latest Nvidia drivers: 525.78.01 and Cuda version 12.
Hi @ChuckPa,
If you could share 1.29.2.6364 RPM, that would be nice… I’m far than ready to migrate to Ubuntu yet! 
(not that I don’t want, but for now, test will be a lost faster if I can test it on my existing server)
Thanks
PS: am I the only one that is disturbed that you shared Jellyfish videos on a Plex forum ??? 
1 Like
@Franck_Ehret
Please check the signature yourself to verify after downloading: https://anonymfile.com/8ppZN/plexmediaserver-12926364-6d72b0cf6x86-64.rpm
Changed the link to one that doesnt get deleted after first download
$ rpm -v --checksig plexmediaserver-1.29.2.6364-6d72b0cf6.x86_64.rpm
plexmediaserver-1.29.2.6364-6d72b0cf6.x86_64.rpm:
Header V4 RSA/SHA1 Signature, key ID 3adca79d: OK
Header SHA1 digest: OK
V4 RSA/SHA1 Signature, key ID 3adca79d: OK
MD5 digest: OK
I’ve downgraded my nVidia driver to v470.141.03, and the problem persists.
Upgraded Plex to 1.31.0.6654 with no improvement.
Where do I set “nvdexExtraFrames” ? i guess it translates to -extra_hw_frames 9 ??
upgraded to plexmediaserver_1.31.1.6617-cbcbc6f56_amd64.deb but transcoder with hevc (Main 10), yuv420p(tv) issue is still there
example of stream 8bit hevc stream :
Stream #0:0: Video: hevc (Main 10), yuv420p(tv), 1920x1080 [SAR 1:1 DAR 16:9], 23.98 fps, 23.98 tbr, 1k tbn, 23.98 tbc (default)
±----------------------------------------------------------------------------+
| NVIDIA-SMI 515.86.01 Driver Version: 515.86.01 CUDA Version: 11.7 |
|-------------------------------±---------------------±---------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 NVIDIA GeForce … Off | 00000000:01:00.0 Off | N/A |
| 29% 41C P0 N/A / 75W | 527MiB / 4096MiB | 1% Default |
| | | N/A |
±------------------------------±---------------------±---------------------+
±----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| 0 N/A N/A 680478 C …diaserver/Plex Transcoder 523MiB |
±----------------------------------------------------------------------------+
I’m having the same experience with Unraid 6.11.5 / Nvidia 525.85.05.
Downgrading Plex to “plexinc/pms-docker:1.30.0.6486-629d58034” reenables hw encoding.
I would like to test version 1.31.1.6617-cbcbc6f56_amd64, is there a latest/rc “channel” available on docker hub?
what you don’t understand??..move down to PMS 1.29.2.6364…and it will work !!!
for now downgrade to PMS 1.29.2.6364…and it will work !
I would if I could. Doesn’t appear the database is compatible. Server sat “unreachable” for over an hour after downgrading.
@CinaedPhoenix Would you be able to provide me with some more detail on your config please?
I’m running same unraid OS, docker image and nvidia drivers (with a P400) but I am still getting transcode errors using android and Web app.
I’m passing through my GPU correctly as I can see the transcoder PID when running nvidia-smi so I am a bit stumped.
Edit: just tried downgrading to v1.29.2.6364 as well and still no luck 
there was a post you can’t downgrade to that version …
so didn’t try that (yet)
after “downgrade”
Response code=“503” title=“Maintenance” status="Plex Media Server is currently running database migration.
but that cleared and server is up with plexmediaserver_1.29.2.6364-6d72b0cf6_amd64.deb
Did one(1) test and can confirm that it now transcodes a Video: hevc (Main 10), yuv420p(tv), 1920x1080
So nvidia 515 (cuda 11.7) with plexmediaserver_1.29.2.6364 work
(didn’t try with nvidia 520 or 525)
Hi
I’m running a GeForce 1050ti, don’t know if that’s the difference.
Have you configured all 3 docker settings to enable GPU encoding?
(NVIDIA_VISIBLE_DEVICES, NVIDIA_DRIVER_CAPABILITIES and --runtime=nvidia)
I first downgraded to a 1.28… release, don’t remember which one, but that fixed it too, you could try on of those?
Have you verified that other Dockers can use the GPU, i setup folding@home to test.
Best regards
Cinaedphoenix
