A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 9, 12:30 UTC
[CI] minor fixes to pipeline generator and tests (#33151)
Signed-off-by: khluu <khluu000@gmail.com>
Kevin H. Luu committed
Jan 27, 2026 at 17:04 UTC
5d3d6e44e82a8733c83104b83c96bbac88321a96
4 files changed
+18
-16
.buildkite/.pipeline_gen_v2
.buildkite/image_build/image_build.sh
+16
-14
@@ -143,7 +143,7 @@ resolve_parent_commit() {
143
print_bake_config() {
144
echo "--- :page_facing_up: Resolved bake configuration"
145
BAKE_CONFIG_FILE="bake-config-build-${BUILDKITE_BUILD_NUMBER:-local}.json"
146
- docker buildx bake -f "${VLLM_BAKE_FILE}" -f "${CI_HCL_PATH}" --print "${TARGET}" | tee "${BAKE_CONFIG_FILE}" || true
146
+ docker buildx bake -f "${VLLM_BAKE_FILE_PATH}" -f "${CI_HCL_PATH}" --print "${TARGET}" | tee "${BAKE_CONFIG_FILE}" || true
147
echo "Saved bake config to ${BAKE_CONFIG_FILE}"
148
echo "--- :arrow_down: Uploading bake config to Buildkite"
149
buildkite-agent artifact upload "${BAKE_CONFIG_FILE}"
@@ -170,9 +170,9 @@ IMAGE_TAG_LATEST=${8:-} # only used for main branch, optional
170
171
# build config
172
TARGET="test-ci"
173
-CI_HCL_URL="${CI_HCL_URL:-https://raw.githubusercontent.com/vllm-project/ci-infra/main/docker/ci.hcl}"
174
-VLLM_BAKE_FILE="${VLLM_BAKE_FILE:-docker/docker-bake.hcl}"
173
+VLLM_BAKE_FILE_PATH="${VLLM_BAKE_FILE_PATH:-docker/docker-bake.hcl}"
174
BUILDER_NAME="${BUILDER_NAME:-vllm-builder}"
175
+CI_HCL_URL="${CI_HCL_URL:-https://raw.githubusercontent.com/vllm-project/ci-infra/main/docker/ci.hcl}"
176
CI_HCL_PATH="/tmp/ci.hcl"
177
BUILDKIT_SOCKET="/run/buildkit/buildkitd.sock"
178
@@ -180,9 +180,8 @@ prepare_cache_tags
180
ecr_login
181
182
# Environment info (for docs and human readers)
183
-# CI_HCL_URL - URL to ci.hcl (default: from ci-infra main branch)
183
# VLLM_CI_BRANCH - ci-infra branch to use (default: main)
185
-# VLLM_BAKE_FILE - Path to vLLM's bake file (default: docker/docker-bake.hcl)
184
+# VLLM_BAKE_FILE_PATH - Path to vLLM's bake file (default: docker/docker-bake.hcl)
185
# BUILDER_NAME - Name for buildx builder (default: vllm-builder)
186
#
187
# Build configuration (exported as environment variables for bake):
@@ -211,10 +210,9 @@ echo "IMAGE_TAG_LATEST: ${IMAGE_TAG_LATEST}"
210
# print build configuration
211
echo "--- :mag: Build configuration"
212
echo "TARGET: ${TARGET}"
214
-echo "CI HCL URL: ${CI_HCL_URL}"
215
-echo "vLLM bake file: ${VLLM_BAKE_FILE}"
213
+echo "vLLM bake file: ${VLLM_BAKE_FILE_PATH}"
214
echo "BUILDER_NAME: ${BUILDER_NAME}"
217
-echo "CI_HCL_PATH: ${CI_HCL_PATH}"
215
+echo "CI_HCL_URL: ${CI_HCL_URL}"
216
echo "BUILDKIT_SOCKET: ${BUILDKIT_SOCKET}"
217
218
echo "--- :mag: Cache tags"
@@ -227,11 +225,11 @@ check_and_skip_if_image_exists
225
226
echo "--- :docker: Setting up Docker buildx bake"
227
echo "Target: ${TARGET}"
230
-echo "CI HCL URL: ${CI_HCL_URL}"
231
-echo "vLLM bake file: ${VLLM_BAKE_FILE}"
228
+echo "vLLM bake file: ${VLLM_BAKE_FILE_PATH}"
229
+echo "CI HCL path: ${CI_HCL_PATH}"
230
233
-if [[ ! -f "${VLLM_BAKE_FILE}" ]]; then
234
- echo "Error: vLLM bake file not found at ${VLLM_BAKE_FILE}"
231
+if [[ ! -f "${VLLM_BAKE_FILE_PATH}" ]]; then
232
+ echo "Error: vLLM bake file not found at ${VLLM_BAKE_FILE_PATH}"
233
echo "Make sure you're running from the vLLM repository root"
234
exit 1
235
fi
@@ -240,15 +238,19 @@ echo "--- :arrow_down: Downloading ci.hcl"
238
curl -sSfL -o "${CI_HCL_PATH}" "${CI_HCL_URL}"
239
echo "Downloaded to ${CI_HCL_PATH}"
240
241
+if [[ ! -f "${CI_HCL_PATH}" ]]; then
242
+ echo "Error: ci.hcl not found at ${CI_HCL_PATH}"
243
+ exit 1
244
+fi
245
+
246
setup_buildx_builder
247
245
-# Compute parent commit for cache fallback (if not already set)
248
resolve_parent_commit
249
export PARENT_COMMIT
250
251
print_bake_config
252
253
echo "--- :docker: Building ${TARGET}"
252
-docker --debug buildx bake -f "${VLLM_BAKE_FILE}" -f "${CI_HCL_PATH}" --progress plain "${TARGET}"
254
+docker --debug buildx bake -f "${VLLM_BAKE_FILE_PATH}" -f "${CI_HCL_PATH}" --progress plain "${TARGET}"
255
256
echo "--- :white_check_mark: Build complete"
.buildkite/image_build/image_build.yaml
+1
-1
@@ -5,7 +5,7 @@ steps:
5
depends_on: []
6
commands:
7
- if [[ "$BUILDKITE_BRANCH" != "main" ]]; then .buildkite/image_build/image_build.sh $REGISTRY $REPO $BUILDKITE_COMMIT $BRANCH $VLLM_USE_PRECOMPILED $VLLM_MERGE_BASE_COMMIT $IMAGE_TAG; fi
8
- - if [[ "$BUILDKITE_BRANCH" == "main" ]]; then .buildkite/image_build/image_build.sh $REGISTRY $REPO $BUILDKITE_COMMIT $BRANCH $VLLM_USE_PRECOMPILED $VLLM_MERGE_BASE_COMMIT $IMAGE_TAG_LATEST; fi
8
+ - if [[ "$BUILDKITE_BRANCH" == "main" ]]; then .buildkite/image_build/image_build.sh $REGISTRY $REPO $BUILDKITE_COMMIT $BRANCH $VLLM_USE_PRECOMPILED $VLLM_MERGE_BASE_COMMIT $IMAGE_TAG $IMAGE_TAG_LATEST; fi
9
retry:
10
automatic:
11
- exit_status: -1 # Agent was lost
.buildkite/test_areas/distributed.yaml
+1
-1
@@ -180,7 +180,7 @@ steps:
180
- tests/distributed/
181
- tests/examples/offline_inference/data_parallel.py
182
commands:
183
- - ./.buildkite/scripts/run-multi-node-test.sh /vllm-workspace/tests 2 2 public.ecr.aws/q9t5s3a7/vllm-ci-postmerge-repo:0bec63fa317e1fbd62e19b0fc31c43c81bf89077 "VLLM_TEST_SAME_HOST=0 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_same_node.py | grep 'Same node test passed' && NUM_NODES=2 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_node_count.py | grep 'Node count test passed' && python3 ../examples/offline_inference/data_parallel.py -dp=2 -tp=1 --dp-num-nodes=2 --dp-node-rank=0 --dp-master-addr=192.168.10.10 --dp-master-port=12345 --enforce-eager --trust-remote-code && VLLM_MULTI_NODE=1 pytest -v -s distributed/test_multi_node_assignment.py && VLLM_MULTI_NODE=1 pytest -v -s distributed/test_pipeline_parallel.py" "VLLM_TEST_SAME_HOST=0 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_same_node.py | grep 'Same node test passed' && NUM_NODES=2 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_node_count.py | grep 'Node count test passed' && python3 ../examples/offline_inference/data_parallel.py -dp=2 -tp=1 --dp-num-nodes=2 --dp-node-rank=1 --dp-master-addr=192.168.10.10 --dp-master-port=12345 --enforce-eager --trust-remote-code"
183
+ - ./.buildkite/scripts/run-multi-node-test.sh /vllm-workspace/tests 2 2 $IMAGE_TAG "VLLM_TEST_SAME_HOST=0 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_same_node.py | grep 'Same node test passed' && NUM_NODES=2 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_node_count.py | grep 'Node count test passed' && python3 ../examples/offline_inference/data_parallel.py -dp=2 -tp=1 --dp-num-nodes=2 --dp-node-rank=0 --dp-master-addr=192.168.10.10 --dp-master-port=12345 --enforce-eager --trust-remote-code && VLLM_MULTI_NODE=1 pytest -v -s distributed/test_multi_node_assignment.py && VLLM_MULTI_NODE=1 pytest -v -s distributed/test_pipeline_parallel.py" "VLLM_TEST_SAME_HOST=0 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_same_node.py | grep 'Same node test passed' && NUM_NODES=2 torchrun --nnodes 2 --nproc-per-node=2 --rdzv_backend=c10d --rdzv_endpoint=192.168.10.10 distributed/test_node_count.py | grep 'Node count test passed' && python3 ../examples/offline_inference/data_parallel.py -dp=2 -tp=1 --dp-num-nodes=2 --dp-node-rank=1 --dp-master-addr=192.168.10.10 --dp-master-port=12345 --enforce-eager --trust-remote-code"
184
185
- label: Distributed NixlConnector PD accuracy (4 GPUs)
186
timeout_in_minutes: 30