Merge pull request #100 from runpod/dj/containers-v1

feat: containers v1

Author: max4c

.github/actions/docker-setup/action.yml

@@ -13,7 +13,7 @@ outputs:
     value: ${{ steps.build-type.outputs.is_production }}
   release-suffix:
     description: 'Release suffix for non-production builds'
-    value: ${{ steps.env-vars.outputs.release }}
+    value: ${{ steps.env-vars.outputs.release_suffix }}
 
 runs:
   using: 'composite'
@@ -32,10 +32,8 @@ runs:
         username: ${{ inputs.dockerhub-username }}
         password: ${{ inputs.dockerhub-token }}
 
-    - name: Set up Docker Buildx
-      uses: useblacksmith/[email protected]
-      with:
-        setup-only: true
+    - name: Set up Docker Build
+      uses: useblacksmith/setup-docker-builder@v1
 
     - name: Set up QEMU
       uses: docker/setup-qemu-action@v3
@@ -52,10 +50,10 @@ runs:
 
     - name: Set Environment Variables
       id: env-vars
+      if: github.ref != 'refs/heads/main' && github.ref != 'refs/heads/dj/containers-v1'
       shell: bash
       run: |
-        if [[ "${{ github.ref }}" != "refs/heads/main" ]]; then
-          BRANCH_NAME=$(echo ${GITHUB_REF##refs/heads/} | sed 's/\//-/g')
-          echo "RELEASE=dev-${BRANCH_NAME}" >> $GITHUB_ENV
-          echo "release=dev-${BRANCH_NAME}" >> $GITHUB_OUTPUT
-        fi 
+        BRANCH_NAME=$(echo ${GITHUB_REF##refs/heads/} | sed 's/\//-/g')
+        SUFFIX="-dev-${BRANCH_NAME}"
+        echo "RELEASE=${SUFFIX}" >> $GITHUB_ENV
+        echo "release_suffix=${SUFFIX}" >> $GITHUB_OUTPUT 

.github/workflows/base.yml

@@ -35,14 +35,13 @@ jobs:
         uses: docker/bake-action@v6
         env:
           BUILDX_BAKE_ENTITLEMENTS_FS: 0
+          RELEASE_SUFFIX: ${{ steps.setup.outputs.release-suffix }}
         with:
           source: .
           files: |
             official-templates/shared/versions.hcl
             official-templates/base/docker-bake.hcl
           push: true
-          set: |
-            ${{ steps.setup.outputs.is-production != 'true' && format('*.args.RELEASE_VERSION={0}', steps.setup.outputs.release-suffix) || '' }}
 
   build-pytorch:
     needs: build-base
@@ -77,11 +76,10 @@ jobs:
         uses: docker/bake-action@v6
         env:
           BUILDX_BAKE_ENTITLEMENTS_FS: 0
+          RELEASE_SUFFIX: ${{ steps.setup.outputs.release-suffix }}
         with:
           source: .
           files: |
             official-templates/shared/versions.hcl
             official-templates/pytorch/docker-bake.hcl
           push: true
-          set: |
-            ${{ steps.setup.outputs.is-production != 'true' && format('*.args.RELEASE_VERSION={0}', steps.setup.outputs.release-suffix) || '' }}

.github/workflows/rocm.yml

@@ -5,8 +5,6 @@ on:
     paths:
       - ".github/workflows/rocm.yml"
       - "official-templates/rocm/**"
-      - "official-templates/shared/**"
-      - "official-templates/base/Dockerfile"
   workflow_dispatch:
 
 permissions:
@@ -33,11 +31,10 @@ jobs:
         uses: docker/bake-action@v6
         env:
           BUILDX_BAKE_ENTITLEMENTS_FS: 0
+          RELEASE_SUFFIX: ${{ steps.setup.outputs.release-suffix }}
         with:
           source: .
           files: |
             official-templates/shared/versions.hcl
             official-templates/rocm/docker-bake.hcl
           push: true
-          set: |
-            ${{ steps.setup.outputs.is-production != 'true' && format('*.args.RELEASE_VERSION={0}', steps.setup.outputs.release-suffix) || '' }}

README.md

@@ -8,21 +8,21 @@ This repository contains the Dockerfiles for the Runpod containers used for our
 
 The following dependencies are required for all images for Runpod platform functionality.
 
-- `nginx` - Required for proxying ports to the user.
-- `openssh-server` - Required for SSH access to the container.
-- `jupyterlab` - Required for JupyterLab access to the container.
+- `nginx`: Required for proxying ports to the user.
+- `openssh-server`: Required for SSH access to the container.
+- `jupyterlab`: Required for JupyterLab access to the container.
 
 ### README
 
-Every container folder has its own README.md file, this file is displayed on the Docker Hub and the README section of the template on the Runpod website. Additionally, if the container opens a port other than 8888 that is passed through the proxy and the service is not running yet, the README will be displayed to the user to guide them.
+Every container folder has its own README.md file, displayed on Docker Hub and in the template section on the Runpod website. When containers open ports other than 8888 through the proxy, the README guides users while services are starting up.
 
 ## Building Containers
 
-This repository is powered by Docker Buildx and uses [bake files](https://docs.docker.com/build/bake/) to manage builds.
+This repository uses Docker Buildx with [bake files](https://docs.docker.com/build/bake/) to manage builds.
 
 ### Using the Bake Script
 
-`./bake.sh` automatically combines shared version definitions with template-specific bake files. 
+`./bake.sh` automatically combines shared version definitions with template specific bake files. 
 
 Use it like this:
 
@@ -39,4 +39,4 @@ Use it like this:
 
 ### Version Definitions
 
-Version compatibility and build targets for CUDA, Ubuntu, and PyTorch is centralized in `official-templates/shared/versions.hcl`. This file is automatically included when building with the `bake.sh` script. When adding new versions or changing compatibility, modify this file.
+Version compatibility and build targets for CUDA, Ubuntu, and PyTorch are centralized in `official-templates/shared/versions.hcl`. This file is automatically included when building with the `bake.sh` script. When adding new versions or changing compatibility, modify this file.

container-template/start.sh

@@ -1,5 +1,5 @@
 #!/bin/bash
-set -e  # Exit the script if any statement returns a non-true return value
+set -e # Exit the script if any statement returns a non-true return value
 
 # ---------------------------------------------------------------------------- #
 #                          Function Definitions                                #
@@ -29,7 +29,7 @@ setup_ssh() {
         echo "$PUBLIC_KEY" >> ~/.ssh/authorized_keys
         chmod 700 -R ~/.ssh
 
-         if [ ! -f /etc/ssh/ssh_host_rsa_key ]; then
+        if [ ! -f /etc/ssh/ssh_host_rsa_key ]; then
             ssh-keygen -t rsa -f /etc/ssh/ssh_host_rsa_key -q -N ''
             echo "RSA key fingerprint:"
             ssh-keygen -lf /etc/ssh/ssh_host_rsa_key.pub
@@ -66,17 +66,19 @@ setup_ssh() {
 # Export env vars
 export_env_vars() {
     echo "Exporting environment variables..."
-    printenv | grep -E '^RUNPOD_|^PATH=|^_=' | awk -F = '{ print "export " $1 "=\"" $2 "\"" }' >> /etc/rp_environment
-    echo 'source /etc/rp_environment' >> ~/.bashrc
+    printenv | grep -E '^[A-Z_][A-Z0-9_]*=' | grep -v '^PUBLIC_KEY' | awk -F = '{ val = $0; sub(/^[^=]*=/, "", val); print "export " $1 "=\"" val "\"" }' > /etc/rp_environment
+    if ! grep -q 'source /etc/rp_environment' ~/.bashrc; then
+        echo 'source /etc/rp_environment' >> ~/.bashrc
+    fi
 }
 
 # Start jupyter lab
 start_jupyter() {
     if [[ $JUPYTER_PASSWORD ]]; then
         echo "Starting Jupyter Lab..."
-        mkdir -p /workspace && \
-        cd / && \
-        nohup python3.10 -m jupyter lab --allow-root --no-browser --port=8888 --ip=* --FileContentsManager.delete_to_trash=False --ServerApp.terminado_settings='{"shell_command":["/bin/bash"]}' --ServerApp.token=$JUPYTER_PASSWORD --ServerApp.allow_origin=* --ServerApp.preferred_dir=/workspace &> /jupyter.log &
+        mkdir -p /workspace &&
+            cd / &&
+            nohup python3 -m jupyter lab --allow-root --no-browser --port=8888 --ip=* --FileContentsManager.delete_to_trash=False --ServerApp.terminado_settings='{"shell_command":["/bin/bash"]}' --IdentityProvider.token=$JUPYTER_PASSWORD --ServerApp.allow_origin=* --ServerApp.preferred_dir=/workspace &> /jupyter.log &
         echo "Jupyter Lab started"
     fi
 }
@@ -95,8 +97,6 @@ setup_ssh
 start_jupyter
 export_env_vars
 
-execute_script "/post_start.sh" "Running post-start script..."
-
 echo "Start script(s) finished, Pod is ready to use."
 
 sleep infinity

official-templates/base/Dockerfile

@@ -31,22 +31,19 @@ ENV TZ=Etc/UTC
 
 WORKDIR /
 
-# Update and upgrade
-RUN apt-get update --yes && \
-    apt-get upgrade --yes
-
 RUN echo "en_US.UTF-8 UTF-8" > /etc/locale.gen
 
 # Install 
-RUN apt-get install --yes --no-install-recommends \
-    ca-certificates curl dirmngr file git gpg gpg-agent inotify-tools \
-    jq lsof nano nginx openssh-server rsync software-properties-common \
-    sudo tmux unzip vim wget zip build-essential make cmake gfortran ffmpeg \
-    libatlas-base-dev libavcodec-dev libavfilter-dev libavformat-dev libblas-dev \
-    libffi-dev libhdf5-dev libgl1 libjpeg-dev liblapack-dev libpng-dev \
-    libpostproc-dev libsm6 libssl-dev libswscale-dev libtiff-dev \
-    libv4l-dev libx264-dev libxrender-dev libxvidcore-dev \
-    cifs-utils nfs-common zstd
+RUN apt-get update --yes && \
+    apt-get upgrade --yes && \
+    apt-get install --yes --no-install-recommends \
+    build-essential ca-certificates cifs-utils cmake curl dirmngr dnsutils ffmpeg \
+    file gfortran git gpg gpg-agent inetutils-traceroute inotify-tools iputils-ping jq \
+    libatlas-base-dev libavcodec-dev libavfilter-dev libavformat-dev libblas-dev libffi-dev \
+    libgl1 libhdf5-dev libjpeg-dev liblapack-dev libnuma-dev libpng-dev libpostproc-dev \
+    libsm6 libssl-dev libswscale-dev libtiff-dev libv4l-dev libx264-dev libxrender-dev \
+    libxvidcore-dev lsof make mtr nano nfs-common nginx openssh-server rsync slurm-wlm \
+    software-properties-common sudo tmux unzip vim wget zip zstd
 
 # Add the Python PPA
 RUN add-apt-repository ppa:deadsnakes/ppa -y
@@ -57,13 +54,13 @@ RUN apt-get install --yes --no-install-recommends \
     python3.10-dev python3.10-venv python3.10-distutils \
     python3.11-dev python3.11-venv python3.11-distutils \
     python3.12-dev python3.12-venv \
-    python3.13-dev python3.13-venv python-is-python3 && \
+    python3.13-dev python3.13-venv && \
     apt-get autoremove -y && \
     apt-get clean && \
     rm -rf /var/lib/apt/lists/*
 
 # Install pip if we are not ROCm
-RUN if [ -z "${ROCM_PATH}"]; then \
+RUN if [ -z "${ROCM_PATH}" ]; then \
         curl https://bootstrap.pypa.io/get-pip.py -o get-pip.py && \
         python3.9  get-pip.py && \
         python3.10 get-pip.py && \
@@ -74,28 +71,27 @@ RUN if [ -z "${ROCM_PATH}"]; then \
     fi
 
 # ensurepip (it's there, we just installed it^) and install virtualenv
-RUN if [ -z "${ROCM_PATH}"]; then \
+RUN if [ -z "${ROCM_PATH}" ]; then \
         python3.9 -m pip install --upgrade pip virtualenv && \
         python3.10 -m pip install --upgrade pip virtualenv && \
         python3.11 -m pip install --upgrade pip virtualenv && \
         python3.12 -m pip install --upgrade pip virtualenv && \
         python3.13 -m pip install --upgrade pip virtualenv; \
     fi
 
-# Install Jupyter and related packages (always with Python 3.10)
-RUN python3.10 -m pip install --upgrade --no-cache-dir \
+RUN ln -sf /usr/bin/python3.12 /usr/local/bin/python
+RUN ln -sf /usr/local/bin/pip3.12 /usr/local/bin/pip
+RUN ln -sf /usr/local/bin/pip3.12 /usr/local/bin/pip3
+
+COPY --from=ghcr.io/astral-sh/uv:latest /uv /uvx /bin/
+
+# Install Jupyter and related packages
+RUN python -m pip install --upgrade --no-cache-dir \
     jupyterlab \
     ipywidgets \
     jupyter-archive \
     notebook==7.4.2
 
-# Install pip drop-in replacement uv (https://github.com/astral-sh/uv)
-RUN curl -LsSf https://astral.sh/uv/install.sh | sh && \
-    echo 'export PATH="$HOME/.local/bin:$PATH"' >> /etc/profile.d/uv.sh && \
-    chmod +x /etc/profile.d/uv.sh && \
-    export PATH="$HOME/.local/bin:$PATH" && \
-    uv --version
-
 # Install filebrowser
 RUN curl -LsSf https://raw.githubusercontent.com/filebrowser/get/master/get.sh | bash
 
@@ -110,9 +106,8 @@ RUN rm -f /etc/ssh/ssh_host_*
 # Copy the README.md
 COPY README.md /usr/share/nginx/html/README.md
 
-# Start Scripts
+# Start Script
 COPY --from=scripts --chmod=755 start.sh /
-COPY --chmod=755 post_start.sh /
 
 # Welcome Message
 COPY --from=logo runpod.txt /etc/runpod.txt

official-templates/base/README.md

@@ -1,96 +1,42 @@
-<h1>Runpod Base</h1>
+### Runpod Base
 
-**Essential foundation for machine learning workflows.**
+**A lean, flexible starting point for machine learning workflows.**
 
-Runpod Base images are designed to be lightweight but flexible—supporting a wide range of user workloads without overwhelming choice when building your own images. Whether you're prototyping a new ML model, setting up a development environment, or deploying a production workload, these images provide the essential foundation you need to get started fast.
+The Runpod Base images provide a clean, developer friendly environment for everything from quick experiments to production, supporting both GPU and CPU-only workloads. Use them standalone for a preconfigured workspace, or as the foundation for your own images.
 
-**What makes these essential:**
-- **Ready-to-use Jupyter Environment** — Both Notebook and JupyterLab with widgets and extensions (optional - can be disabled by not setting `JUPYTER_PASSWORD`)
-- **Multiple Python Versions** — Python 3.9 through 3.13 installed and ready to use, with 3.10 as the default
-- **Smart Workspace Setup** — Pre-configured directory structure with optimized cache locations for faster package installs
-- **Developer Tools** — SSH access for remote development and NGINX server for web services
-- **ML-Ready Libraries** — Core dependencies for scientific computing, image processing, and machine learning workflows
+### What's included
+- **Multiple Python versions**: 3.9–3.13 preinstalled; 3.10 is the default.
+- **ML ready**: Essential libraries for scientific computing, computer vision, and machine learning, plus SLURM support.
+- **Developer friendly**: SSH server preconfigured for seamless remote development and debugging.
+- **Smart workspace**: Optimized directory structure and package caches for faster dependency installation.
+- **Performance tuned**: Environment variables and cache strategies optimized for faster builds and execution.
+- **Jupyter ready (optional)**: Notebook and JupyterLab with widgets/extensions; enable by setting `JUPYTER_PASSWORD` (omit to disable).
 
-**Choose your combination:**
-- **Ubuntu versions:** 20.04, 22.04, or 24.04
-- **CUDA versions:** 12.4.1 through 12.9.0 for GPU acceleration
-- **Optimized Performance:** Efficiently configured environment variables and cache directories
+### Available configurations
+- **Ubuntu**: 22.04 (Jammy) and 24.04 (Noble)
+- **CUDA**: 12.8.0, 12.8.1, 12.9.0, and 13.0.0
 
-Perfect for immediate development and experimenting, or as the foundation for more specialized containers.
-
-*Need something more specialized? Check out our other templates in the `official-templates` directory for ROCm support, PyTorch, and other ML frameworks.*
+Need something more specialized? Explore the templates in `official-templates` for ROCm, PyTorch, and more.
 
 <div class="base-images">
 
 ## Generated Images
 
-### Operating Systems:
-- Ubuntu 20.04:
-    - `runpod/base:0.7.0`
-    - `runpod/base:0.7.0-ubuntu2004`
-- Ubuntu 22.04:
-    - `runpod/base:0.7.0-ubuntu2204`
-    - `runpod/base:0.7.0-jammy`
-- Ubuntu 24.04:
-    - `runpod/base:0.7.0-ubuntu2404`
-    - `runpod/base:0.7.0-noble`
+### Base Images (CPU-Only, No GPU Drivers):
+- Ubuntu 22.04: `runpod/base:1.0.2-ubuntu2204`
+- Ubuntu 24.04: `runpod/base:1.0.2-ubuntu2404`
 
-### CUDA Versions:
-- 12.4.1:
-    - Ubuntu 20.04:
-        - `runpod/base:0.7.0-cuda1241`
-        - `runpod/base:0.7.0-cuda1241-ubuntu2004`
-        - `runpod/base:0.7.0-focal-cuda1241`
-    - Ubuntu 22.04:
-        - `runpod/base:0.7.0-cuda1241`
-        - `runpod/base:0.7.0-cuda1241-ubuntu2204`
-        - `runpod/base:0.7.0-jammy-cuda1241`
-- 12.5.1:
-    - Ubuntu 20.04:
-        - `runpod/base:0.7.0-cuda1251`
-        - `runpod/base:0.7.0-cuda1251-ubuntu2004`
-        - `runpod/base:0.7.0-focal-cuda1251`
-    - Ubuntu 22.04:
-        - `runpod/base:0.7.0-cuda1251`
-        - `runpod/base:0.7.0-cuda1251-ubuntu2204`
-        - `runpod/base:0.7.0-jammy-cuda1251`
-- 12.6.3:
-    - Ubuntu 20.04:
-        - `runpod/base:0.7.0-cuda1263`
-        - `runpod/base:0.7.0-cuda1263-ubuntu2004`
-        - `runpod/base:0.7.0-focal-cuda1263`
-    - Ubuntu 22.04:
-        - `runpod/base:0.7.0-cuda1263`
-        - `runpod/base:0.7.0-cuda1263-ubuntu2204`
-        - `runpod/base:0.7.0-jammy-cuda1263`
-    - Ubuntu 24.04:
-        - `runpod/base:0.7.0-cuda1263`
-        - `runpod/base:0.7.0-cuda1263-ubuntu2404`
-        - `runpod/base:0.7.0-noble-cuda1263`
+### CUDA Images (GPU Required) by Version:
+- 12.8.0:
+    - Ubuntu 22.04: `runpod/base:1.0.2-cuda1280-ubuntu2204`
+    - Ubuntu 24.04: `runpod/base:1.0.2-cuda1280-ubuntu2404`
 - 12.8.1:
-    - Ubuntu 20.04:
-        - `runpod/base:0.7.0-cuda1281`
-        - `runpod/base:0.7.0-cuda1281-ubuntu2004`
-        - `runpod/base:0.7.0-focal-cuda1281`
-    - Ubuntu 22.04:
-        - `runpod/base:0.7.0-cuda1281`
-        - `runpod/base:0.7.0-cuda1281-ubuntu2204`
-        - `runpod/base:0.7.0-jammy-cuda1281`
-    - Ubuntu 24.04:
-        - `runpod/base:0.7.0-cuda1281`
-        - `runpod/base:0.7.0-cuda1281-ubuntu2404`
-        - `runpod/base:0.7.0-noble-cuda1281`
+    - Ubuntu 22.04: `runpod/base:1.0.2-cuda1281-ubuntu2204`
+    - Ubuntu 24.04: `runpod/base:1.0.2-cuda1281-ubuntu2404`
 - 12.9.0:
-    - Ubuntu 20.04:
-        - `runpod/base:0.7.0-cuda1290`
-        - `runpod/base:0.7.0-cuda1290-ubuntu2004`
-        - `runpod/base:0.7.0-focal-cuda1290`
-    - Ubuntu 22.04:
-        - `runpod/base:0.7.0-cuda1290`
-        - `runpod/base:0.7.0-cuda1290-ubuntu2204`
-        - `runpod/base:0.7.0-jammy-cuda1290`
-    - Ubuntu 24.04:
-        - `runpod/base:0.7.0-cuda1290`
-        - `runpod/base:0.7.0-cuda1290-ubuntu2404`
-        - `runpod/base:0.7.0-noble-cuda1290`
+    - Ubuntu 22.04: `runpod/base:1.0.2-cuda1290-ubuntu2204`
+    - Ubuntu 24.04: `runpod/base:1.0.2-cuda1290-ubuntu2404`
+- 13.0.0:
+    - Ubuntu 22.04: `runpod/base:1.0.2-cuda1300-ubuntu2204`
+    - Ubuntu 24.04: `runpod/base:1.0.2-cuda1300-ubuntu2404`
 </div>