Auto conversion of /watch files fails: Cannot load libcuda.so.1 about handbrake-nvenc-docker HOT 4 CLOSED

thank you for this report, I will look into it, although I have to add that I am officially not supporting unRAID

from handbrake-nvenc-docker.

I sadly cannot reproduce this issue, it works for me using a 1080ti and the latest image from Docker Hub

my conversion log

[11:09:27] encavcodecInit: H.265 (Nvidia NVENC)
[11:09:27] encavcodec: encoding at rc=vbr, multipass=fullres, 20.00
[11:09:27] encavcodec: encoding with stored aspect 1/1
[11:09:27] sync: first pts video is 0
[11:09:27] sync: Chapter 1 at frame 1 time 0
[11:09:27] sync: first pts audio 0x1 is 0
[11:10:32] reader: done. 1 scr changes
[11:10:33] work: average encoding speed for job is 186.046875 fps
[11:10:33] aac-decoder done: 17413 frames, 0 decoder errors
[11:10:33] h264-decoder done: 12128 frames, 0 decoder errors
[11:10:33] sync: got 12128 frames, 12129 expected
[11:10:33] sync: framerate min 30.000 fps, max 30.000 fps, avg 30.000 fps
[11:10:33] mux: track 0, 12128 frames, 467774485 bytes, 9254.46 kbps, fifo 512
[11:10:33] mux: track 1, 17414 frames, 7921494 bytes, 156.72 kbps, fifo 1024
[11:10:33] Finished work at: Sat Jun 25 11:10:33 2022

[11:10:33] libhb: work result = 0

Encode done!
HandBrake has exited.

could you send me your preset that you are using?

from handbrake-nvenc-docker.

I figured this out, as I expected it was user error. I did all the unRAID stuff to tell the container to use the GPU but I skipped the step that tells docker to let it use the GPU, oops! I was wondering why it was working in the GUI and not in the terminal and it turns out that selecting the preset it in the GUI didn't break because it defaulted to the encoders that were available but specifying the preset from the command line did not. I thought it was working because it encoded but did not look closely enough, because it was definitely not. All is working now though!

In case anyone finds this using unRAID I had to pass --runtime=nvidia as an extra parameter to the container(this is the bit I missed) and set the NVIDIA_VISIBLE_DEVICES env variable to my GPU device ID. This was all that was required in addition to the env variables set in the docker file.

from handbrake-nvenc-docker.

nice glad to hear!

from handbrake-nvenc-docker.