Dockers are not loading with default runtime as nvidia on ver 3.4.0

[santoshyadav30]

On meta-tegra-holoscan ver 3.4.0, while running the docker with runtime as nvidia below issue is observed:

# docker run nvidia/cuda:12.5.0-base-ubuntu22.04
docker: Error response from daemon: failed to create task for container: failed to create shim task: OCI runtime create failed: unable to retrieve OCI runtime error (open /run/containerd/io.containerd.runtime.v2.task/moby/b8b4bfdaee52f84768ede42be6aef3040e0aefa379689e7cfd8c8413de90dfa0/log.json: no such file or directory): nvidia-container-runtime did not terminate successfully: exit status 127: unknown.
ERRO[0014] error waiting for container: context canceled 

Running nvidia-container-runtime also results in issue:
# nvidia-container-runtime
nvidia-container-runtime: symbol lookup error: nvidia-container-runtime: undefined symbol: cuDeviceGet

Below are the details of the nvidia container utilities/libraries:

# nvidia-container-cli info
NVRM version:   555.42.02
CUDA version:   12.5
Device Index:   0
Device Minor:   0
Model:          NVIDIA RTX 6000 Ada Generation
Brand:          NvidiaRTX
GPU UUID:       GPU-d752896a-1bec-0be2-4347-b168d0f7595a
Bus Location:   00000005:09:00.0
Architecture:   8.9

libnvidia-container version : 1.11.0
nvidia-container-toolkit version : 1.14.4