guidellm/Containerfile at main · vllm-project/guidellm · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
ARG BASE_IMAGE=quay.io/fedora/python-313-minimal:latest

# Use a multi-stage build to create a lightweight production image
FROM $BASE_IMAGE as builder

# release: take the last version and add a post if build iteration
# candidate: increment to next minor, add 'rc' with build iteration
# nightly: increment to next minor, add 'a' with build iteration
# alpha: increment to next minor, add 'a' with build iteration
# dev: increment to next minor, add 'dev' with build iteration
ARG GUIDELLM_BUILD_TYPE=dev

# Extra dependencies to install
# all: install all extras
# recommended: install recommended extras
ARG GUIDELLM_BUILD_EXTRAS=all

# Switch to root for installing packages
USER root

# Install uv in a temporary venv
RUN /usr/bin/python3 -m venv /tmp/uv \
    && /tmp/uv/bin/pip install --no-cache-dir -U uv \
    && ln -s /tmp/uv/bin/uv /usr/local/bin/uv

# Install build dependencies
RUN --mount=type=cache,sharing=locked,target=/var/cache/dnf \
    dnf install -y git

# Set correct build type for versioning
# Configure uv for building guidellm
ENV GUIDELLM_BUILD_TYPE=$GUIDELLM_BUILD_TYPE \
    VIRTUAL_ENV=/opt/app-root \
    UV_PROJECT="/src" \
    UV_LINK_MODE="copy" \
    UV_NO_DEV="1" \
    UV_NO_EDITABLE="1" \
    UV_FROZEN="1" \
    UV_CACHE_DIR="/tmp/uv_cache"

# Sync initial environment
RUN --mount=type=cache,target=$UV_CACHE_DIR \
    --mount=type=bind,source=uv.lock,target=$UV_PROJECT/uv.lock,relabel=shared \
    --mount=type=bind,source=pyproject.toml,target=$UV_PROJECT/pyproject.toml,relabel=shared \
    uv sync --active --no-install-project --extra $GUIDELLM_BUILD_EXTRAS

# Copy repository files
# Do this as late as possible to leverage layer caching
COPY / $UV_PROJECT

# Install guidellm
RUN --mount=type=cache,target=$UV_CACHE_DIR \
    uv sync --active --extra $GUIDELLM_BUILD_EXTRAS

# Prod image
FROM $BASE_IMAGE

# Switch to root for installing packages
USER root

# Install some helpful utilities and deps
RUN --mount=type=cache,sharing=locked,target=/var/cache/dnf \
    dnf install -y --setopt=install_weak_deps=False \
        vi tar rsync ffmpeg-free

# Switch back to unpriv user
# Root group for k8s
USER 1001:0

# Add guidellm bin to PATH
# Argument defaults can be set with GUIDELLM_<ARG>
ENV HOME="/home/guidellm" \
    GUIDELLM_OUTPUT_DIR="/results"

# Create the user home dir
WORKDIR $HOME

# Ensure that the user home dir can be used by any user
# (OpenShift Pods can't use the cache otherwise)
RUN chgrp -R 0 "$HOME" && chmod -R g=u "$HOME"

# Create a volume for results
VOLUME /results

# Metadata
LABEL io.k8s.display-name="GuideLLM" \
      org.opencontainers.image.description="GuideLLM Performance Benchmarking Container" \
      org.opencontainers.image.source="https://github.com/vllm-project/guidellm" \
      org.opencontainers.image.documentation="https://blog.vllm.ai/guidellm/stable" \
      org.opencontainers.image.license="Apache-2.0"

# Copy the virtual environment from the builder stage
# Do this as late as possible to leverage layer caching
COPY --chown=1001:0 --from=builder /opt/app-root /opt/app-root

ENTRYPOINT [ "/opt/app-root/bin/guidellm" ]
CMD [ "benchmark", "run" ]