[Currency] SGLang 0.5.7 v1 (#5589)

* [Currency] SGLang 0.5.7 v1

* fix sglang build

* increase file limit

* increase file limit

* test true

* revert toml

Author: sirutBuasai

release_images_general.yml

@@ -119,7 +119,7 @@ release_images:
       private_registry: False
   9:
     framework: "sglang"
-    version: "0.5.6"
+    version: "0.5.7"
     arch_type: "x86"
     customer_type: "sagemaker"
     general:

sglang/buildspec-sm.yml

@@ -2,7 +2,7 @@ account_id: &ACCOUNT_ID <set-$ACCOUNT_ID-in-environment>
 prod_account_id: &PROD_ACCOUNT_ID 763104351884
 region: &REGION <set-$REGION-in-environment>
 framework: &FRAMEWORK sglang
-version: &VERSION "0.5.6"
+version: &VERSION "0.5.7"
 short_version: &SHORT_VERSION "0.5"
 arch_type: &ARCH_TYPE x86_64
 autopatch_build: "False"
@@ -36,12 +36,12 @@ images:
     <<: *BUILD_REPOSITORY
     context:
       <<: *BUILD_CONTEXT
-    image_size_baseline: 26000
+    image_size_baseline: 32000
     device_type: &DEVICE_TYPE gpu
     cuda_version: &CUDA_VERSION cu129
     python_version: &DOCKER_PYTHON_VERSION py3
     tag_python_version: &TAG_PYTHON_VERSION py312
-    os_version: &OS_VERSION ubuntu22.04
+    os_version: &OS_VERSION ubuntu24.04
     tag: !join [ *VERSION, "-", *DEVICE_TYPE, "-", *TAG_PYTHON_VERSION, "-", *CUDA_VERSION, "-", *OS_VERSION, "-sagemaker" ]
     latest_release_tag: !join [ *VERSION, "-", *DEVICE_TYPE, "-", *TAG_PYTHON_VERSION, "-", *CUDA_VERSION, "-", *OS_VERSION, "-sagemaker" ]
     skip_build: "False"

sglang/x86_64/gpu/Dockerfile

@@ -1,4 +1,4 @@
-FROM lmsysorg/sglang:v0.5.6-cu129-amd64 AS base
+FROM lmsysorg/sglang:v0.5.7-cu129-amd64 AS base
 
 # ====================================================
 # ====================== common ======================
@@ -11,16 +11,16 @@ LABEL maintainer="Amazon AI"
 LABEL dlc_major_version="1"
 
 ENV DEBIAN_FRONTEND=noninteractive \
-    LANG=C.UTF-8 \
-    LC_ALL=C.UTF-8 \
-    DLC_CONTAINER_TYPE=general \
-    # Python won’t try to write .pyc or .pyo files on the import of source modules
-    # Force stdin, stdout and stderr to be totally unbuffered. Good for logging
-    PYTHONDONTWRITEBYTECODE=1 \
-    PYTHONUNBUFFERED=1 \
-    PYTHONIOENCODING=UTF-8 \
-    LD_LIBRARY_PATH="/usr/local/lib:/opt/amazon/ofi-nccl/lib/x86_64-linux-gnu:/opt/amazon/openmpi/lib:/opt/amazon/efa/lib:/usr/local/cuda/lib64:${LD_LIBRARY_PATH}" \
-    PATH="/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda/bin:${PATH}"
+  LANG=C.UTF-8 \
+  LC_ALL=C.UTF-8 \
+  DLC_CONTAINER_TYPE=general \
+  # Python won’t try to write .pyc or .pyo files on the import of source modules
+  # Force stdin, stdout and stderr to be totally unbuffered. Good for logging
+  PYTHONDONTWRITEBYTECODE=1 \
+  PYTHONUNBUFFERED=1 \
+  PYTHONIOENCODING=UTF-8 \
+  LD_LIBRARY_PATH="/usr/local/lib:/opt/amazon/ofi-nccl/lib/x86_64-linux-gnu:/opt/amazon/openmpi/lib:/opt/amazon/efa/lib:/usr/local/cuda/lib64:${LD_LIBRARY_PATH}" \
+  PATH="/opt/amazon/openmpi/bin:/opt/amazon/efa/bin:/usr/local/cuda/bin:${PATH}"
 
 WORKDIR /
 
@@ -35,53 +35,44 @@ RUN chmod +x /usr/local/bin/deep_learning_container.py \
  && chmod +x /usr/local/bin/bash_telemetry.sh \
  && chmod +x /usr/local/bin/start_cuda_compat.sh
 
-# Install cuda compat
-# ====================
-# RUN apt-get update \
-#  && apt-get -y upgrade --only-upgrade systemd \
-#  && apt-get install -y --allow-change-held-packages --no-install-recommends \
-#     cuda-compat-12-9 \
-#  && rm -rf /var/lib/apt/lists/* \
-#  && apt-get clean
-
 # Install EFA and remove vulnerable nvjpeg
 # =========================================
 RUN bash install_efa.sh ${EFA_VERSION} \
- && rm install_efa.sh \
- && mkdir -p /tmp/nvjpeg \
- && cd /tmp/nvjpeg \
- # latest cu12 libnvjpeg available is cu124
- && wget https://developer.download.nvidia.com/compute/cuda/redist/libnvjpeg/linux-x86_64/libnvjpeg-linux-x86_64-12.4.0.76-archive.tar.xz \
- && tar -xvf libnvjpeg-linux-x86_64-12.4.0.76-archive.tar.xz \
- && rm -rf /usr/local/cuda/targets/x86_64-linux/lib/libnvjpeg* \
- && rm -rf /usr/local/cuda/targets/x86_64-linux/include/nvjpeg.h \
- && cp libnvjpeg-linux-x86_64-12.4.0.76-archive/lib/libnvjpeg* /usr/local/cuda/targets/x86_64-linux/lib/ \
- && cp libnvjpeg-linux-x86_64-12.4.0.76-archive/include/* /usr/local/cuda/targets/x86_64-linux/include/ \
- && rm -rf /tmp/nvjpeg \
- # create symlink for python
- && rm -rf /usr/bin/python \
- && ln -s /usr/bin/python3 /usr/bin/python \
- # remove cuobjdump and nvdisasm
- && rm -rf /usr/local/cuda/bin/cuobjdump* \
- && rm -rf /usr/local/cuda/bin/nvdisasm*
+  && rm install_efa.sh \
+  && mkdir -p /tmp/nvjpeg \
+  && cd /tmp/nvjpeg \
+  # latest cu12 libnvjpeg available is cu124
+  && wget https://developer.download.nvidia.com/compute/cuda/redist/libnvjpeg/linux-x86_64/libnvjpeg-linux-x86_64-12.4.0.76-archive.tar.xz \
+  && tar -xvf libnvjpeg-linux-x86_64-12.4.0.76-archive.tar.xz \
+  && rm -rf /usr/local/cuda/targets/x86_64-linux/lib/libnvjpeg* \
+  && rm -rf /usr/local/cuda/targets/x86_64-linux/include/nvjpeg.h \
+  && cp libnvjpeg-linux-x86_64-12.4.0.76-archive/lib/libnvjpeg* /usr/local/cuda/targets/x86_64-linux/lib/ \
+  && cp libnvjpeg-linux-x86_64-12.4.0.76-archive/include/* /usr/local/cuda/targets/x86_64-linux/include/ \
+  && rm -rf /tmp/nvjpeg \
+  # create symlink for python
+  && rm -rf /usr/bin/python \
+  && ln -s /usr/bin/python3 /usr/bin/python \
+  # remove cuobjdump and nvdisasm
+  && rm -rf /usr/local/cuda/bin/cuobjdump* \
+  && rm -rf /usr/local/cuda/bin/nvdisasm*
 
 # Run OSS compliance script
 # ==========================
-RUN echo 'source /usr/local/bin/bash_telemetry.sh' >> /etc/bash.bashrc  \
- # OSS compliance - use Python zipfile instead of unzip
- && HOME_DIR=/root \
- && curl -o ${HOME_DIR}/oss_compliance.zip https://aws-dlinfra-utilities.s3.amazonaws.com/oss_compliance.zip \
- && python3 -c "import zipfile, os; zipfile.ZipFile('/root/oss_compliance.zip').extractall('/root/'); os.remove('/root/oss_compliance.zip')" \
- && cp ${HOME_DIR}/oss_compliance/test/testOSSCompliance /usr/local/bin/testOSSCompliance \
- && chmod +x /usr/local/bin/testOSSCompliance \
- && chmod +x ${HOME_DIR}/oss_compliance/generate_oss_compliance.sh \
- && ${HOME_DIR}/oss_compliance/generate_oss_compliance.sh ${HOME_DIR} ${PYTHON} \
- # clean up
- && rm -rf ${HOME_DIR}/oss_compliance* \
- && rm -rf /tmp/tmp* \
- && rm -rf /tmp/uv* \
- && rm -rf /var/lib/apt/lists/* \
- && rm -rf /root/.cache | true
+RUN echo 'source /usr/local/bin/bash_telemetry.sh' >> /etc/bash.bashrc \
+  # OSS compliance - use Python zipfile instead of unzip
+  && HOME_DIR=/root \
+  && curl -o ${HOME_DIR}/oss_compliance.zip https://aws-dlinfra-utilities.s3.amazonaws.com/oss_compliance.zip \
+  && python3 -c "import zipfile, os; zipfile.ZipFile('/root/oss_compliance.zip').extractall('/root/'); os.remove('/root/oss_compliance.zip')" \
+  && cp ${HOME_DIR}/oss_compliance/test/testOSSCompliance /usr/local/bin/testOSSCompliance \
+  && chmod +x /usr/local/bin/testOSSCompliance \
+  && chmod +x ${HOME_DIR}/oss_compliance/generate_oss_compliance.sh \
+  && ${HOME_DIR}/oss_compliance/generate_oss_compliance.sh ${HOME_DIR} ${PYTHON} \
+  # clean
+  && rm -rf ${HOME_DIR}/oss_compliance* \
+  && rm -rf /tmp/tmp* \
+  && rm -rf /tmp/uv* \
+  && rm -rf /var/lib/apt/lists/* \
+  && rm -rf /root/.cache | true
 
 # =======================================================
 # ====================== sagemaker ======================
@@ -90,9 +81,9 @@ RUN echo 'source /usr/local/bin/bash_telemetry.sh' >> /etc/bash.bashrc  \
 FROM base AS sglang-sagemaker
 
 RUN dpkg -l | grep -E "cuda|nvidia|libnv" | awk '{print $2}' | xargs apt-mark hold \
- && apt-get update \
- && apt-get upgrade -y \
- && apt-get clean
+  && apt-get update \
+  && apt-get upgrade -y \
+  && apt-get clean
 
 RUN rm -rf /tmp/*