Could not run sample deviceQuery application inside docker

Accelerated Computing NGC GPU Cloud Container: CUDA
1

Hi, I am trying to run sample deviceQuery application inside docker container on xavier but getting error as follows:
unresolvable R_AARCH64_ADR_PREL_PG_HI21 relocation against symbol __STACK_CHK_GUARD@@GLIBC_2.17.

I have os l4t version -R31
and ubuntu 18.04

Also I am facing this problem while following this given document : NVIDIA Container Runtime on Jetson · NVIDIA/nvidia-docker Wiki · GitHub

I also have my own compute engine which I have containerized with all the dependencies install inside docker using shell script and using base image as ubuntu:18.04.
cuda version: 10.1

steps:
run my image using command:

sudo docker run -it --privileged --device=/dev/nvhost-ctrl --device=/dev/nvhost-ctrl-gpu --device=/dev/nvhost-prof-gpu --device=/dev/nvmap --device=/dev/nvhost-gpu --device=/dev/nvhost-as-gpu -v /usr/lib/aarch64-linux-gnu:/usr/lib/aarch64-linux-gnu -v /usr/local:/usr/local -v /usr/lib:/usr/lib docker890/computejetsonmodule:1.0.2 /bin/bash

go inside docker container:
sudo docker exec -it containerID /bin/bash

run deviceQuery sample application

This gives me error as follows:

./deviceQuery
./deviceQuery Starting…

CUDA Device Query (Runtime API) version (CUDART static linking)

cudaGetDeviceCount returned 35
→ CUDA driver version is insufficient for CUDA runtime version
Result = FAIL

Can somebody please help me with this?

2

Did anyone find a solution to the above?

3

Note that you are not using NGC here. This particular forum is related to using NGC.

With respect to this question:

“Also I am facing this problem while following this given document : NVIDIA Container Runtime on Jetson · NVIDIA/nvidia-docker Wiki · GitHub”

“CUDA driver version is insufficient for CUDA runtime version” usually means just what it says.

The driver version is too old for the CUDA runtime version that you are trying to use (10.1)

The latest jetpack at this time (4.3) appears to include CUDA 10.0 (10.0.326):

https://docs.nvidia.com/jetson/jetpack/release-notes/

Your Jetson install appears to be based on an older Jetpack (R31 OS instead of R32) but even if you were on the latest Jetpack, I wouldn’t expect it to support CUDA 10.1

I think you may have better luck if you build your container image around CUDA 10.0

With respect to the remainder, its possible you built your container incorrectly.