In rhel8.4 GPU-Driver-Container image why install elfutils-libelf

[lengrongfu]

The template below is mostly useful for bug reports and support questions. Feel free to remove anything which doesn't apply to you and add more information where it makes sense.

Important Note: NVIDIA AI Enterprise customers can get support from NVIDIA Enterprise support. Please open a case here.

1. Quick Debug Information

• OS/Version(e.g. RHEL8.6, Ubuntu22.04): rhel8.4
• Kernel Version: 4.18.0-305.el8.x86_64
• Container Runtime Type/Version(e.g. Containerd, CRI-O, Docker): Containerd
• K8s Flavor/Version(e.g. K8s, OCP, Rancher, GKE, EKS): K8s
• GPU Operator Version: 1.23.1

2. Issue or feature description

Briefly explain the issue in terms of expected behavior and current behavior.

in the nvcr.io/nvidia/driver:525.105.17-rhel8.4 image having elfutils-libelf lib, but code need install it.

https://github.com/NVIDIA/GPU-Driver-Container/blob/368849ac5e27fa0e4382b4702e786fb9868ecbaf/rhel8/nvidia-driver#L86C28-L86C43

[shivamerla]

@lengrongfu looks like only elfutils-libelf is present but not elfutils-devel and we are installing both which are required for build.

[lengrongfu]

yes, we need to install elfutils-devel, but nvidia-driver script execution dnf install -q -y elfutils-libelf.$DRIVER_ARCH elfutils-libelf-devel.$DRIVER_ARCH, will install them at the same time, It will require that the version of elfutils-devel is the same as that of elfutils-libelf, but the version of elfutils-libelf has exceeded the version of elfutils-devel provided by redhat.

In our scenario, we are installing offline and have our own rpm warehouse of rhel8.4 iso, but the elfutils-libelf-devel version provided by rhel8.4 is elfutils-libelf-0.182-3.el8.x86_64.rpm, but nvcr.io/nvidia/driver:525.105.17-rhel8.4 The elfutils-libelf version provided is elfutils-libelf-0.187-4.el8.x86_64. There is a mismatch between the two versions.