From c2287bba56e6de513329a2ba233564b142fc3000 Mon Sep 17 00:00:00 2001 From: Dillon Cullinan Date: Thu, 30 Oct 2025 14:22:50 -0400 Subject: [PATCH 1/7] Use docker proxy for SGLANG Signed-off-by: Dillon Cullinan --- .github/actions/docker-build/action.yml | 4 +++- container/Dockerfile.sglang | 4 ++-- 2 files changed, 5 insertions(+), 3 deletions(-) diff --git a/.github/actions/docker-build/action.yml b/.github/actions/docker-build/action.yml index ed4617bd9c..b3ceb822cc 100644 --- a/.github/actions/docker-build/action.yml +++ b/.github/actions/docker-build/action.yml @@ -94,6 +94,7 @@ runs: AWS_ACCESS_KEY_ID: ${{ inputs.aws_access_key_id }} AWS_SECRET_ACCESS_KEY: ${{ inputs.aws_secret_access_key }} PLATFORM: ${{ inputs.platform }} + ECR_HOSTNAME: ${{ inputs.aws_account_id }}.dkr.ecr.${{ inputs.aws_default_region }}.amazonaws.com run: | # Determine image tag if [ -n "${{ inputs.image_tag }}" ]; then @@ -129,7 +130,8 @@ runs: --platform ${{ inputs.platform }} \ --use-sccache \ --sccache-bucket "$SCCACHE_S3_BUCKET" \ - --sccache-region "$AWS_DEFAULT_REGION" $EXTRA_ARGS + --sccache-region "$AWS_DEFAULT_REGION" \ + --build-arg DOCKER_PROXY=${ECR_HOSTNAME}/dockerhub/ $EXTRA_ARGS BUILD_END_TIME=$(date -u +%Y-%m-%dT%H:%M:%SZ) echo "🕐 Build ended at: ${BUILD_END_TIME}" diff --git a/container/Dockerfile.sglang b/container/Dockerfile.sglang index 976b5f551b..1f18f05591 100644 --- a/container/Dockerfile.sglang +++ b/container/Dockerfile.sglang @@ -39,8 +39,8 @@ FROM ${DYNAMO_BASE_IMAGE} AS dynamo_base # - Develop or debug framework-level components # - Create custom builds with specific optimization flags # - -FROM nvidia/cuda:${CUDA_VERSION}-cudnn-devel-ubuntu24.04 AS framework +ARG DOCKER_PROXY +FROM ${DOCKER_PROXY}nvidia/cuda:${CUDA_VERSION}-cudnn-devel-ubuntu24.04 AS framework # Declare all ARGs ARG BUILD_TYPE=all From 443e77b484095a7e4298f77315960ec227cb32dd Mon Sep 17 00:00:00 2001 From: Dillon Cullinan Date: Thu, 30 Oct 2025 14:50:07 -0400 Subject: [PATCH 2/7] Debug container runtime Signed-off-by: Dillon Cullinan --- .github/actions/pytest/action.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/actions/pytest/action.yml b/.github/actions/pytest/action.yml index 0037129a5e..3c2ec6ecf0 100644 --- a/.github/actions/pytest/action.yml +++ b/.github/actions/pytest/action.yml @@ -53,7 +53,7 @@ runs: run: | # Run pytest with detailed output and JUnit XML set +e # Don't exit on test failures - + sudo cat /etc/docker/daemon.json docker run --runtime=nvidia --rm --gpus all -w /workspace \ --cpus=${NUM_CPUS} \ --network host \ From cf711008e41de0da32a5145d8961092d7c90fe0a Mon Sep 17 00:00:00 2001 From: Dillon Cullinan Date: Thu, 30 Oct 2025 14:50:58 -0400 Subject: [PATCH 3/7] Revert debug Signed-off-by: Dillon Cullinan --- .github/actions/pytest/action.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/actions/pytest/action.yml b/.github/actions/pytest/action.yml index 0037129a5e..3c2ec6ecf0 100644 --- a/.github/actions/pytest/action.yml +++ b/.github/actions/pytest/action.yml @@ -53,7 +53,7 @@ runs: run: | # Run pytest with detailed output and JUnit XML set +e # Don't exit on test failures - + sudo cat /etc/docker/daemon.json docker run --runtime=nvidia --rm --gpus all -w /workspace \ --cpus=${NUM_CPUS} \ --network host \ From 1aaa9bc44deebea5a6f4863f923ea7f312f7de0c Mon Sep 17 00:00:00 2001 From: Dillon Cullinan Date: Thu, 30 Oct 2025 14:51:30 -0400 Subject: [PATCH 4/7] Revert debug Signed-off-by: Dillon Cullinan --- .github/actions/pytest/action.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/actions/pytest/action.yml b/.github/actions/pytest/action.yml index 3c2ec6ecf0..0037129a5e 100644 --- a/.github/actions/pytest/action.yml +++ b/.github/actions/pytest/action.yml @@ -53,7 +53,7 @@ runs: run: | # Run pytest with detailed output and JUnit XML set +e # Don't exit on test failures - sudo cat /etc/docker/daemon.json + docker run --runtime=nvidia --rm --gpus all -w /workspace \ --cpus=${NUM_CPUS} \ --network host \ From feb01872eff6916c6eb44da27489cf1e7a1061a7 Mon Sep 17 00:00:00 2001 From: Dillon Cullinan Date: Thu, 30 Oct 2025 14:52:33 -0400 Subject: [PATCH 5/7] Revert test Signed-off-by: Dillon Cullinan --- .github/actions/pytest/action.yml | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/.github/actions/pytest/action.yml b/.github/actions/pytest/action.yml index 3c2ec6ecf0..0037129a5e 100644 --- a/.github/actions/pytest/action.yml +++ b/.github/actions/pytest/action.yml @@ -53,7 +53,7 @@ runs: run: | # Run pytest with detailed output and JUnit XML set +e # Don't exit on test failures - sudo cat /etc/docker/daemon.json + docker run --runtime=nvidia --rm --gpus all -w /workspace \ --cpus=${NUM_CPUS} \ --network host \ From 07dcf7919741b4990d812dfe4570432f95119f39 Mon Sep 17 00:00:00 2001 From: Tushar Sharma Date: Thu, 30 Oct 2025 12:23:40 -0700 Subject: [PATCH 6/7] Use nvcr.io image instead of dockerhub image Signed-off-by: Tushar Sharma --- .github/actions/docker-build/action.yml | 4 +--- container/Dockerfile.sglang | 2 +- 2 files changed, 2 insertions(+), 4 deletions(-) diff --git a/.github/actions/docker-build/action.yml b/.github/actions/docker-build/action.yml index b3ceb822cc..ed4617bd9c 100644 --- a/.github/actions/docker-build/action.yml +++ b/.github/actions/docker-build/action.yml @@ -94,7 +94,6 @@ runs: AWS_ACCESS_KEY_ID: ${{ inputs.aws_access_key_id }} AWS_SECRET_ACCESS_KEY: ${{ inputs.aws_secret_access_key }} PLATFORM: ${{ inputs.platform }} - ECR_HOSTNAME: ${{ inputs.aws_account_id }}.dkr.ecr.${{ inputs.aws_default_region }}.amazonaws.com run: | # Determine image tag if [ -n "${{ inputs.image_tag }}" ]; then @@ -130,8 +129,7 @@ runs: --platform ${{ inputs.platform }} \ --use-sccache \ --sccache-bucket "$SCCACHE_S3_BUCKET" \ - --sccache-region "$AWS_DEFAULT_REGION" \ - --build-arg DOCKER_PROXY=${ECR_HOSTNAME}/dockerhub/ $EXTRA_ARGS + --sccache-region "$AWS_DEFAULT_REGION" $EXTRA_ARGS BUILD_END_TIME=$(date -u +%Y-%m-%dT%H:%M:%SZ) echo "🕐 Build ended at: ${BUILD_END_TIME}" diff --git a/container/Dockerfile.sglang b/container/Dockerfile.sglang index 1f18f05591..39078cfc74 100644 --- a/container/Dockerfile.sglang +++ b/container/Dockerfile.sglang @@ -40,7 +40,7 @@ FROM ${DYNAMO_BASE_IMAGE} AS dynamo_base # - Create custom builds with specific optimization flags # ARG DOCKER_PROXY -FROM ${DOCKER_PROXY}nvidia/cuda:${CUDA_VERSION}-cudnn-devel-ubuntu24.04 AS framework +FROM nvcr.io/nvidia/cuda:${CUDA_VERSION}-cudnn-devel-ubuntu24.04 AS framework # Declare all ARGs ARG BUILD_TYPE=all From 1463f4b7790bb32655352eefe6048164df1b2a54 Mon Sep 17 00:00:00 2001 From: Tushar Sharma Date: Thu, 30 Oct 2025 12:43:58 -0700 Subject: [PATCH 7/7] Remove DOCKER_PROXY ARG Signed-off-by: Tushar Sharma --- container/Dockerfile.sglang | 1 - 1 file changed, 1 deletion(-) diff --git a/container/Dockerfile.sglang b/container/Dockerfile.sglang index 39078cfc74..3a036a5524 100644 --- a/container/Dockerfile.sglang +++ b/container/Dockerfile.sglang @@ -39,7 +39,6 @@ FROM ${DYNAMO_BASE_IMAGE} AS dynamo_base # - Develop or debug framework-level components # - Create custom builds with specific optimization flags # -ARG DOCKER_PROXY FROM nvcr.io/nvidia/cuda:${CUDA_VERSION}-cudnn-devel-ubuntu24.04 AS framework # Declare all ARGs