[CI] Use uv-installed standalone Python in vllm-base (drop deadsnakes)

The vllm-base stage installs python3.12 via the deadsnakes PPA, which
requires reaching launchpad.net (REST API) or ppa.launchpadcontent.net
(CDN). On the release-v2 build agents, both are unreachable
(Connection timed out to 185.125.190.80), causing the regular Ubuntu
22.04 release images to hang or fail at the apt-get update step.

Switch vllm-base to the same approach the `base` stage already uses:
uv installs a standalone python3.12 (via python-build-standalone,
hosted on astral.sh CDN) into /opt/venv, and we symlink it to
/usr/bin/python3. No Launchpad dependency.

Side benefits:
- Consistent python install path between `base` and `vllm-base`.
- Removes the software-properties-common / add-apt-repository step.
- Drops unused DEADSNAKES_MIRROR_URL / DEADSNAKES_GPGKEY_URL /
  GET_PIP_URL stage args.

Test branch: khluu/release-v0.20.1-uv-python (off releases/v0.20.1).

Signed-off-by: khluu <khluu000@gmail.com>

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

.buildkite/release-pipeline.yaml

@@ -121,7 +121,7 @@ steps:
           queue: cpu_queue_release
         commands:
           - "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
-          - "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=13.0.2 --build-arg torch_cuda_arch_list=\"${CUDA_ARCH_X86}\" --build-arg INSTALL_KV_CONNECTORS=true --build-arg BUILD_BASE_IMAGE=nvidia/cuda:13.0.2-devel-ubuntu22.04 --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m) --target vllm-openai --progress plain -f docker/Dockerfile ."
+          - "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=13.0.2 --build-arg torch_cuda_arch_list=\"${CUDA_ARCH_X86}\" --build-arg INSTALL_KV_CONNECTORS=true --build-arg DEADSNAKES_MIRROR_URL=https://ppa.launchpadcontent.net/deadsnakes/ppa/ubuntu --build-arg DEADSNAKES_GPGKEY_URL='https://keyserver.ubuntu.com/pks/lookup?op=get&search=0xF23C5A6CF475977595C89F51BA6932366A755776' --build-arg BUILD_BASE_IMAGE=nvidia/cuda:13.0.2-devel-ubuntu22.04 --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m) --target vllm-openai --progress plain -f docker/Dockerfile ."
           - "docker push public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m)"
           # re-tag to default image tag and push, just in case arm64 build fails
           - "docker tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m) public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT"
@@ -134,7 +134,7 @@ steps:
           queue: arm64_cpu_queue_release
         commands:
           - "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
-          - "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=13.0.2 --build-arg torch_cuda_arch_list=\"${CUDA_ARCH_AARCH64}\" --build-arg INSTALL_KV_CONNECTORS=true --build-arg BUILD_BASE_IMAGE=nvidia/cuda:13.0.2-devel-ubuntu22.04 --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m) --target vllm-openai --progress plain -f docker/Dockerfile ."
+          - "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=13.0.2 --build-arg torch_cuda_arch_list=\"${CUDA_ARCH_AARCH64}\" --build-arg INSTALL_KV_CONNECTORS=true --build-arg DEADSNAKES_MIRROR_URL=https://ppa.launchpadcontent.net/deadsnakes/ppa/ubuntu --build-arg DEADSNAKES_GPGKEY_URL='https://keyserver.ubuntu.com/pks/lookup?op=get&search=0xF23C5A6CF475977595C89F51BA6932366A755776' --build-arg BUILD_BASE_IMAGE=nvidia/cuda:13.0.2-devel-ubuntu22.04 --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m) --target vllm-openai --progress plain -f docker/Dockerfile ."
           - "docker push public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m)"
 
       - label: "Build release image - x86_64 - CUDA 12.9"
@@ -144,7 +144,7 @@ steps:
           queue: cpu_queue_release
         commands:
           - "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
-          - "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.9.1 --build-arg torch_cuda_arch_list=\"${CUDA_ARCH_X86_CU129}\" --build-arg INSTALL_KV_CONNECTORS=true --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m)-cu129 --target vllm-openai --progress plain -f docker/Dockerfile ."
+          - "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.9.1 --build-arg torch_cuda_arch_list=\"${CUDA_ARCH_X86_CU129}\" --build-arg INSTALL_KV_CONNECTORS=true --build-arg DEADSNAKES_MIRROR_URL=https://ppa.launchpadcontent.net/deadsnakes/ppa/ubuntu --build-arg DEADSNAKES_GPGKEY_URL='https://keyserver.ubuntu.com/pks/lookup?op=get&search=0xF23C5A6CF475977595C89F51BA6932366A755776' --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m)-cu129 --target vllm-openai --progress plain -f docker/Dockerfile ."
           - "docker push public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m)-cu129"
           # re-tag to default image tag and push, just in case arm64 build fails
           - "docker tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m)-cu129 public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-cu129"
@@ -157,7 +157,7 @@ steps:
           queue: arm64_cpu_queue_release
         commands:
           - "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
-          - "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.9.1 --build-arg torch_cuda_arch_list=\"${CUDA_ARCH_AARCH64_CU129}\" --build-arg INSTALL_KV_CONNECTORS=true --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m)-cu129 --target vllm-openai --progress plain -f docker/Dockerfile ."
+          - "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.9.1 --build-arg torch_cuda_arch_list=\"${CUDA_ARCH_AARCH64_CU129}\" --build-arg INSTALL_KV_CONNECTORS=true --build-arg DEADSNAKES_MIRROR_URL=https://ppa.launchpadcontent.net/deadsnakes/ppa/ubuntu --build-arg DEADSNAKES_GPGKEY_URL='https://keyserver.ubuntu.com/pks/lookup?op=get&search=0xF23C5A6CF475977595C89F51BA6932366A755776' --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m)-cu129 --target vllm-openai --progress plain -f docker/Dockerfile ."
           - "docker push public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT-$(uname -m)-cu129"
 
       - label: "Build release image - x86_64 - CUDA 13.0 - Ubuntu 24.04"

docker/Dockerfile

@@ -487,12 +487,10 @@ WORKDIR /vllm-workspace
 RUN PYTHON_VERSION_STR=$(echo ${PYTHON_VERSION} | sed 's/\.//g') && \
     echo "export PYTHON_VERSION_STR=${PYTHON_VERSION_STR}" >> /etc/environment
 
-# Install Python (via uv / python-build-standalone) and system dependencies.
-# This replaces the deadsnakes PPA, removing the build-time dependency on
-# Launchpad and matching how the build-stage (`base`) installs Python.
-# python-build-standalone bundles dev headers, the venv module, and
-# python3-config, so the python3.X-dev / python3.X-venv apt packages
-# are not needed.
+# Install system dependencies and Python via uv-managed standalone build.
+# This avoids the deadsnakes PPA / Launchpad, which is not always reachable
+# from CI build agents on Ubuntu 22.04 base images. Ubuntu 24.04 base images
+# could use distro python3.12 directly, but using uv keeps both paths uniform.
 RUN apt-get update -y \
     && apt-get install -y --no-install-recommends \
         curl \
@@ -507,12 +505,11 @@ RUN apt-get update -y \
     && $HOME/.local/bin/uv venv /opt/venv --python ${PYTHON_VERSION} \
     && rm -f /usr/bin/python3 /usr/bin/python3-config /usr/bin/pip \
     && ln -s /opt/venv/bin/python3 /usr/bin/python3 \
-    && ln -s /opt/venv/bin/python${PYTHON_VERSION} /usr/bin/python${PYTHON_VERSION} \
     && ln -s /opt/venv/bin/python3-config /usr/bin/python3-config \
     && ln -s /opt/venv/bin/pip /usr/bin/pip \
     && python3 --version && python3 -m pip --version
 
-# Activate virtual environment and add uv to PATH
+# Activate venv: ensures `uv`, `pip`, `python` resolve to the /opt/venv install
 ENV PATH="/opt/venv/bin:/root/.local/bin:$PATH"
 ENV VIRTUAL_ENV="/opt/venv"
 
@@ -540,6 +537,9 @@ RUN CUDA_VERSION_DASH=$(echo $CUDA_VERSION | cut -d. -f1,2 | tr '.' '-') && \
     apt-get install -y --no-install-recommends --allow-change-held-packages libnccl-dev=${NCCL_VER} libnccl2=${NCCL_VER} && \
     rm -rf /var/lib/apt/lists/*
 
+# Install uv for faster pip installs
+RUN python3 -m pip install uv
+
 # Environment for uv
 ENV UV_HTTP_TIMEOUT=500
 ENV UV_INDEX_STRATEGY="unsafe-best-match"
@@ -729,7 +729,7 @@ ENV HF_XET_HIGH_PERFORMANCE 1
 ENV HF_HUB_DOWNLOAD_TIMEOUT 60
 
 # Copy in the v1 package for testing (it isn't distributed yet)
-COPY vllm/v1 /opt/venv/lib/python${PYTHON_VERSION}/site-packages/vllm/v1
+COPY vllm/v1 /usr/local/lib/python${PYTHON_VERSION}/dist-packages/vllm/v1
 
 # Source code is used in the `python_only_compile.sh` test
 # We hide it inside `src/` so that this source code