Add AL2023 Dockerfile equivalent to Deadline Cloud SMF worker AMI

[leongdl]

Add AL2023 Dockerfile equivalent to Deadline Cloud SMF worker AMI

Replicates the ~480 packages installed on top of base AL2023 Minimal AMI on Deadline Cloud service-managed fleet workers. Organized into logical layers: system tools, build toolchain, X11/Mesa/OpenGL, image libs, networking/NFS, Git, Python 3.11, Docker/containerd, and misc libraries.

Useful for building and testing conda packages locally with matching GLIBC and library versions.

Fixes:

What was the problem/requirement? (What/Why)

Building and testing conda packages for Deadline Cloud SMF workers requires matching the exact system library versions (GLIBC, Mesa, X11, etc.) present on the worker AMI. Without a local equivalent, developers either test directly on workers (slow feedback loop) or use a generic AL2023 image that's missing ~480 packages the worker has, leading to "works locally, breaks on worker" issues.

What was the solution? (How)

Diffed the 768 installed packages on a live SMF worker against the 289-package AL2023 Minimal AMI baseline (April 2026) and created a Dockerfile that installs the delta, organized into logical layers:

1. Core system tools and utilities
2. Build toolchain (GCC, binutils, kernel headers)
3. X11, Mesa/OpenGL, Wayland — headless rendering deps
4. Image/media libraries (libjpeg, libpng, libtiff, libwebp)
5. Networking, NFS, SSSD, security
6. Git and Perl
7. Python 3.11
8. Docker + containerd
9. Misc libraries (boost, jemalloc, tbb, etc.)

EC2-specific packages (SSM agent, CloudWatch agent, NVIDIA toolkit, kernel modules) are intentionally excluded with comments explaining why and how to add GPU support if needed.

What is the impact of this change?

No impact on existing samples. This adds a new Dockerfile under containers/al2023-deadline/ that developers can use to replicate the SMF worker environment locally for conda package development and testing.

How was this change tested?

• Built the image locally with docker build:

  docker build -f containers/al2023-deadline/Dockerfile.worker-equivalent -t al2023-deadline-worker:latest .
  

• Build completed successfully, all 10 layers installed without errors.
• Verified the resulting image contains the expected packages (GCC, Mesa, X11, Python 3.11, etc.).
• Used the image as a base for building and testing Blender 5.1 conda packages with rattler-build, confirming GLIBC and library compatibility with the real SMF worker environment.

Was this change documented?

The Dockerfile includes inline comments explaining each layer, the rationale for excluded packages, and instructions for adding NVIDIA GPU support.

---

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.