Skip to content

revert: restore Dockerfile.vllm and build/run scripts to August 28 vers #2892

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
keivenchang merged 2 commits into from
Sep 5, 2025
Merged

revert: restore Dockerfile.vllm and build/run scripts to August 28 vers #2892

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
509b77e
revert: restore Dockerfile.vllm and build/run scripts to August 28 be…
keivenchang Sep 4, 2025
e0857d1
style: clean up unused variables and formatting
keivenchang Sep 5, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
134 changes: 122 additions & 12 deletions container/Dockerfile.vllm
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -280,21 +280,55 @@ RUN --mount=type=bind,source=./container/launch_message.txt,target=/workspace/la
ENTRYPOINT ["/opt/nvidia/nvidia_entrypoint.sh"]
CMD []

#######################################
########## Local Development #######
#######################################
#######################################################################
########## DEVELOPMENT TARGETS FEATURE MATRIX #########################
#######################################################################
# Feature │ local-dev Target │ dev Target
# ─────────────────────┼─────────────────────┼─────────────────────
# Purpose │ Dev Container │ Command-line with
# │ plugin use only │ run.sh script
# ─────────────────────┼─────────────────────┼─────────────────────
# Default User │ ubuntu user │ root user
# ─────────────────────┼─────────────────────┼─────────────────────
# User Setup │ Full ubuntu user │ No user setup
# │ with UID/GID │
# │ mapping │
# ─────────────────────┼─────────────────────┼─────────────────────
# Permissions │ ubuntu user with │ Root-level
# │ sudo privileges │ permissions
# ─────────────────────┼─────────────────────┼─────────────────────
# Home Directory │ /home/ubuntu │ /root
# ─────────────────────┼─────────────────────┼─────────────────────
# Working Directory │ /home/ubuntu/dynamo │ /workspace
# ─────────────────────┼─────────────────────┼─────────────────────
# Rust Toolchain │ User's home │ System locations
# │ (~/.rustup, │ (/usr/local/rustup,
# │ ~/.cargo) │ /usr/local/cargo)
# ─────────────────────┼─────────────────────┼─────────────────────
# Python Environment │ User-owned venv │ System location
# │ │ (/opt/dynamo/venv)
# ─────────────────────┼─────────────────────┼─────────────────────
# File Permissions │ User-level with │ Root-level
# │ proper ownership │ permissions
# ─────────────────────┼─────────────────────┼─────────────────────
# Compatibility │ MS Plug-in: Dev │ Backward compatibility
# │ Container workflow │ with existing
# │ │ workflows
#
# PURPOSE: Local development
#
# This stage adds development tools, utilities, and dependencies specifically
# needed for:
# - Local development and debugging
# - vscode/cursor development
# USAGE GUIDELINES:
# • Use local-dev: VS Code/Cursor Dev Container plugin only
# • Use dev: run.sh script for command-line development


#######################################################################
########## Development (Dev Container only) ###########################
#######################################################################
#
# Use this stage when you need a full development environment with additional
# tooling beyond the base runtime image.
# This stage is for Dev Container plug-in use only.
# It provides a local development environment with extra tools and dependencies
# not present in the base runtime image.

FROM runtime AS dev
FROM runtime AS local-dev

# Don't want ubuntu to be editable, just change uid and gid.
ENV USERNAME=ubuntu
Expand Down Expand Up @@ -395,3 +429,79 @@ RUN mkdir -p /home/$USERNAME/.cache/

ENTRYPOINT ["/opt/nvidia/nvidia_entrypoint.sh"]
CMD []


###########################################################
########## Development (run.sh, runs as root user) ########
###########################################################
#
# PURPOSE: Local development environment for use with run.sh (not Dev Container plug-in)
#
# This stage runs as root and provides:
# - Development tools and utilities for local debugging
# - Support for vscode/cursor development outside the Dev Container plug-in
#
# Use this stage if you need a full-featured development environment with extra tools,
# but do not use it with the Dev Container plug-in.

FROM runtime AS dev

# Don't want ubuntu to be editable, just change uid and gid.
ARG WORKSPACE_DIR=/workspace

# Install utilities as root
RUN apt-get update -y && \
apt-get install -y --no-install-recommends \
# Install utilities
nvtop \
wget \
tmux \
vim \
git \
openssh-client \
iproute2 \
rsync \
zip \
unzip \
htop \
# Build Dependencies
autoconf \
automake \
cmake \
libtool \
meson \
net-tools \
pybind11-dev \
# Rust build dependencies
clang \
libclang-dev \
protobuf-compiler && \
rm -rf /var/lib/apt/lists/*

COPY --from=runtime /usr/local/bin /usr/local/bin

# Set workspace directory variable
ENV WORKSPACE_DIR=${WORKSPACE_DIR} \
DYNAMO_HOME=${WORKSPACE_DIR} \
RUSTUP_HOME=/usr/local/rustup \
CARGO_HOME=/usr/local/cargo \
CARGO_TARGET_DIR=/workspace/target \
VIRTUAL_ENV=/opt/dynamo/venv \
PATH=/usr/local/cargo/bin:$PATH

COPY --from=dynamo_base /usr/local/rustup /usr/local/rustup
COPY --from=dynamo_base /usr/local/cargo /usr/local/cargo

# This is a slow operation (~40s on my cpu)
# Much better than chown -R $USERNAME:$USERNAME /opt/dynamo/venv (~10min on my cpu)
COPY --from=runtime ${VIRTUAL_ENV} ${VIRTUAL_ENV}

# so we can use maturin develop
RUN uv pip install maturin[patchelf]

# Make sure to sync this with the one specified on README.md.
# This is a generic PYTHONPATH which works for all the frameworks, so some paths may not be relevant for this particular framework.
ENV PYTHONPATH=${WORKSPACE_DIR}/components/metrics/src:${WORKSPACE_DIR}/components/frontend/src:${WORKSPACE_DIR}/components/planner/src:${WORKSPACE_DIR}/components/backends/mocker/src:${WORKSPACE_DIR}/components/backends/trtllm/src:${WORKSPACE_DIR}/components/backends/vllm/src:${WORKSPACE_DIR}/components/backends/sglang/src:${WORKSPACE_DIR}/components/backends/llama_cpp/src

ENTRYPOINT ["/opt/nvidia/nvidia_entrypoint.sh"]
CMD []
39 changes: 2 additions & 37 deletions container/build.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,8 +24,6 @@ set -e
TAG=
RUN_PREFIX=
PLATFORM=linux/amd64
USER_UID=
USER_GID=

# Get short commit hash
commit_id=$(git rev-parse --short HEAD)
Expand Down Expand Up @@ -245,22 +243,6 @@ get_options() {
missing_requirement "$1"
fi
;;
--uid)
if [ "$2" ]; then
USER_UID="$2"
shift
else
missing_requirement "$1"
fi
;;
--gid)
if [ "$2" ]; then
USER_GID="$2"
shift
else
missing_requirement "$1"
fi
;;
--dry-run)
RUN_PREFIX="echo"
echo ""
Expand Down Expand Up @@ -437,8 +419,6 @@ show_help() {
echo " [--cache-from cache location to start from]"
echo " [--cache-to location where to cache the build output]"
echo " [--tag tag for image]"
echo " [--uid user ID for dev target (default: current user)]"
echo " [--gid group ID for dev target (default: current group)]"
echo " [--no-cache disable docker build cache]"
echo " [--dry-run print docker commands without running]"
echo " [--build-context name=path to add build context]"
Expand All @@ -463,14 +443,6 @@ error() {

get_options "$@"

# Validate UID/GID flags are only used with dev target
if [ -n "$USER_UID" ] || [ -n "$USER_GID" ]; then
if [[ "$TARGET" != "dev" ]]; then
echo "⚠️ Warning: --uid and --gid flags are only effective with --target dev"
echo " Current target: ${TARGET:-}"
fi
fi

# Automatically set ARCH and ARCH_ALT if PLATFORM is linux/arm64
ARCH="amd64"
if [[ "$PLATFORM" == *"linux/arm64"* ]]; then
Expand All @@ -492,15 +464,8 @@ fi
# Add NIXL_REF as a build argument
BUILD_ARGS+=" --build-arg NIXL_REF=${NIXL_REF} "

if [[ $TARGET == "dev" ]]; then
# Use provided UID/GID or default to current user
if [ -z "$USER_UID" ]; then
USER_UID=$(id -u)
fi
if [ -z "$USER_GID" ]; then
USER_GID=$(id -g)
fi
BUILD_ARGS+=" --build-arg USER_UID=$USER_UID --build-arg USER_GID=$USER_GID "
if [[ $TARGET == "local-dev" ]]; then
BUILD_ARGS+=" --build-arg USER_UID=$(id -u) --build-arg USER_GID=$(id -g) "
fi

# BUILD DEV IMAGE
Expand Down
84 changes: 25 additions & 59 deletions container/run.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -31,14 +31,12 @@ DEFAULT_FRAMEWORK=VLLM
SOURCE_DIR=$(dirname "$(readlink -f "$0")")

IMAGE=
TARGET="dev"
HF_CACHE=
DEFAULT_HF_CACHE=${SOURCE_DIR}/.cache/huggingface
GPUS="all"
PRIVILEGED=
VOLUME_MOUNTS=
MOUNT_WORKSPACE=
DEV_MODE=
ENVIRONMENT_VARIABLES=
REMAINING_ARGS=
INTERACTIVE=
Expand Down Expand Up @@ -164,7 +162,6 @@ get_options() {
--mount-workspace)
MOUNT_WORKSPACE=TRUE
;;

--use-nixl-gds)
USE_NIXL_GDS=TRUE
;;
Expand Down Expand Up @@ -230,11 +227,32 @@ get_options() {
ENTRYPOINT_STRING="--entrypoint ${ENTRYPOINT}"
fi

if [ -n "$MOUNT_WORKSPACE" ]; then
VOLUME_MOUNTS+=" -v ${SOURCE_DIR}/..:/workspace "
VOLUME_MOUNTS+=" -v /tmp:/tmp "
VOLUME_MOUNTS+=" -v /mnt/:/mnt "

if [ -z "$HF_CACHE" ]; then
HF_CACHE=$DEFAULT_HF_CACHE
fi

if [ -z "${PRIVILEGED}" ]; then
PRIVILEGED="TRUE"
fi

ENVIRONMENT_VARIABLES+=" -e HF_TOKEN"

INTERACTIVE=" -it "
fi

if [[ ${HF_CACHE^^} == "NONE" ]]; then
HF_CACHE=
fi

# HF_CACHE mounting will be handled in workspace section
if [ -n "$HF_CACHE" ]; then
mkdir -p "$HF_CACHE"
VOLUME_MOUNTS+=" -v $HF_CACHE:/root/.cache/huggingface"
fi

if [ -z "${PRIVILEGED}" ]; then
PRIVILEGED="FALSE"
Expand All @@ -244,9 +262,9 @@ get_options() {
RM="TRUE"
fi

# Initialize PRIVILEGED_STRING
PRIVILEGED_STRING=""
if [[ ${PRIVILEGED^^} != "FALSE" ]]; then
if [[ ${PRIVILEGED^^} == "FALSE" ]]; then
PRIVILEGED_STRING=""
else
PRIVILEGED_STRING="--privileged"
fi

Expand All @@ -273,23 +291,13 @@ get_options() {
RUNTIME=""
fi

# Auto-enable DEV_MODE for vllm dev images
# TODO(keivenc): Currently only Dockerfile.vllm has proper permissions to run as ubuntu user.
# Other Dockerfiles (trtllm, sglang, etc.) still require root access.
if [[ "$IMAGE" == *"-vllm-dev" ]]; then
DEV_MODE=TRUE
MOUNT_WORKSPACE=TRUE
# Interactive mode is implied when MOUNT_WORKSPACE is TRUE
fi

REMAINING_ARGS=("$@")
}

show_help() {
echo "usage: run.sh"
echo " [--image image]"
echo " [--framework framework one of ${!FRAMEWORKS[*]}]"
echo " [--target target stage to use, default is 'dev']"
echo " [--name name for launched container, default NONE]"
echo " [--privileged whether to launch in privileged mode, default FALSE unless mounting workspace]"
echo " [--dry-run print docker commands without running]"
Expand Down Expand Up @@ -318,48 +326,6 @@ error() {

get_options "$@"

# Process workspace mounting after auto-detection
if [ -n "$MOUNT_WORKSPACE" ]; then
HOME_PATH="/home/ubuntu"

# Common workspace setup
VOLUME_MOUNTS+=" -v $(dirname "${SOURCE_DIR}"):/workspace "
VOLUME_MOUNTS+=" -v /tmp:/tmp "
VOLUME_MOUNTS+=" -v /mnt/:/mnt "
WORKDIR=/workspace
INTERACTIVE=" -it "

# Set default HF_CACHE if not specified
if [ -z "$HF_CACHE" ]; then
HF_CACHE=$DEFAULT_HF_CACHE
fi

# Environment variables for all workspace modes
ENVIRONMENT_VARIABLES+=" -e HF_TOKEN"
ENVIRONMENT_VARIABLES+=" -e GITHUB_TOKEN"
ENVIRONMENT_VARIABLES+=" -e HOME=$HOME_PATH"

# Mount HF_CACHE to user's home cache directory
if [ -n "$HF_CACHE" ]; then
mkdir -p "$HF_CACHE"
VOLUME_MOUNTS+=" -v $HF_CACHE:$HOME_PATH/.cache/huggingface"
fi

if [ -n "$DEV_MODE" ]; then
# Dev Container-specific setup - the Dockerfile handles UID/GID mapping via build args
# This currently only works with Dockerfile.vllm which has proper ubuntu user setup.
echo "Dev Container mode enabled - using ubuntu user with host UID/GID"
# Use ubuntu user (with correct UID/GID baked into image)
PRIVILEGED_STRING+=" --user ubuntu"
else
# Standard workspace mode - enable privileged mode
# TODO(keivenc): Security risk, remove soon. Dockerfiles (trtllm, sglang) still need to run as root.
if [ -z "${PRIVILEGED}" ]; then
PRIVILEGED_STRING="--privileged"
fi
fi
fi

# RUN the image
if [ -z "$RUN_PREFIX" ]; then
set -x
Expand Down
Loading