This repository has been archived by the owner on Jan 22, 2024. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 2k
libnvcuvid.so is missing #1001
Comments
You should check if libnvcuvid is picked up by your host's ldconfig (ldconfig-p | grep cuvid).
|
So this issue has been resolved. Initially we were deploying both CUDA toolkit and drivers. Now we are deploying just the drivers (latest). |
1 task
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
I am trying to build a docker image based off CUDA 10.1 for Ubuntu 16.04.
The application running in the docker container requires libnvcuvid.so. However, this library is not present in the Docker container.
In order to for the application to run successfully I have had to copy the libnvcuvid.so.418.67 library from host machine (which has the CUDA 10.1 toolkit installed) to the /usr/lib/x86_64-linux-gnu/ folder.
I have defined the following environment settings in my Dockerfile
ENV NVIDIA_VISIBLE_DEVICES all
ENV NVIDIA_DRIVER_CAPABILITIES video,compute,utility
Is there a more elegant way to obtain this library (i.e. via an apt package)
uname -a
Linux 5ab363ee00c7 4.15.0-50-generic flag provided but not defined: --build-arg #54~16.04.1-Ubuntu SMP Wed May 8 15:55:19 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux
nvidia-smi -a
docker version
Docker version is 18.09.6
dpkg -l '*nvidia*'
orrpm -qa '*nvidia*'
No packages found
FROM nvidia/cuda:10.1-base-ubuntu16.04
The text was updated successfully, but these errors were encountered: