merge master into branch

Signed-off-by: mattip <matti.picus@gmail.com>

Author: mattip

.buildkite/core.rayci.yml

@@ -292,19 +292,20 @@ steps:
       - "3.12"
       - "3.13"
 
-  # cpp tests
   - label: ":ray: core: cgroup tests"
     tags: core_cpp
     instance_type: medium
     commands:
-      - bazel run //ci/ray_ci:test_in_docker -- //:all //src/... core --only-tags=cgroup --build-type cgroup
-        --privileged --cache-test-results
+      - RAYCI_DISABLE_TEST_DB=1 bazel run //ci/ray_ci:test_in_docker -- //:all //src/ray/common/cgroup2/tests/... core --build-type clang --cache-test-results
+      - docker run --privileged -i --rm --volume /tmp/artifacts:/artifact-mount --shm-size=2.5gb
+        "$${RAYCI_WORK_REPO}":"$${RAYCI_BUILD_ID}"-corebuild /bin/bash
+        "./src/ray/common/cgroup2/integration_tests/sysfs_cgroup_driver_integration_test_entrypoint.sh"
 
   - label: ":ray: core: cpp tests"
     tags: core_cpp
     instance_type: medium
     commands:
-      - bazel run //ci/ray_ci:test_in_docker -- //:all //src/... core --except-tags=cgroup --build-type clang
+      - RAYCI_DISABLE_TEST_DB=1 bazel run //ci/ray_ci:test_in_docker -- //:all //src/... core --except-tags=cgroup --build-type clang
         --cache-test-results --parallelism-per-worker 2
 
   # block on premerge and microcheck
@@ -317,7 +318,7 @@ steps:
     tags: core_cpp
     instance_type: medium
     commands:
-      - bazel run //ci/ray_ci:test_in_docker -- //:all //src/... core --except-tags=cgroup
+      - RAYCI_DISABLE_TEST_DB=1 bazel run //ci/ray_ci:test_in_docker -- //:all //src/... core --except-tags=cgroup
         --build-type asan-clang --cache-test-results --parallelism-per-worker 2
     depends_on:
       - block-core-cpp-sanitizer-tests
@@ -327,7 +328,7 @@ steps:
     tags: core_cpp
     instance_type: large
     commands:
-      - bazel run //ci/ray_ci:test_in_docker -- //:all //src/... core
+      - RAYCI_DISABLE_TEST_DB=1 bazel run //ci/ray_ci:test_in_docker -- //:all //src/... core
         --build-type ubsan --except-tags no_ubsan,cgroup
         --cache-test-results --parallelism-per-worker 2
     depends_on:
@@ -338,27 +339,13 @@ steps:
     tags: core_cpp
     instance_type: medium
     commands:
-      - bazel run //ci/ray_ci:test_in_docker -- //:all //src/... core
+      - RAYCI_DISABLE_TEST_DB=1 bazel run //ci/ray_ci:test_in_docker -- //:all //src/... core
         --build-type tsan-clang --except-tags no_tsan,cgroup
         --cache-test-results --parallelism-per-worker 2
     depends_on:
       - block-core-cpp-sanitizer-tests
       - corebuild
 
-  - label: ":ray: core: flaky cpp tests"
-    key: core_flaky_cpp_tests
-    tags:
-      - python
-      - flaky
-      - skip-on-premerge
-    instance_type: large
-    soft_fail: true
-    commands:
-      - bazel run //ci/ray_ci:test_in_docker -- //:all //src/... core
-        --run-flaky-tests --build-type clang
-    depends_on:
-      - corebuild
-
   - label: ":ray: core: flaky tests"
     key: core_flaky_tests
     tags:

.buildkite/lint.rayci.yml

@@ -35,11 +35,21 @@ steps:
     commands:
       - ./ci/lint/lint.sh pre_commit_pydoclint
 
-  - label: ":lint-roller: lint: {{matrix}}"
+  - label: ":lint-roller: python API: {{matrix}}"
     tags:
       - oss
-      - lint
-      - always
+      - python
+      - dashboard
+      - ray_client
+      - data
+      - serve
+      - ml
+      - tune
+      - train
+      - llm
+      - rllib
+      - rllib_gpu
+      - doc
     key: lint-medium
     instance_type: medium
     depends_on: docbuild

.github/CODEOWNERS

@@ -39,7 +39,7 @@
 
 # Common directory shared by core and the libraries.
 # @edoakes is the czar for now because the pattern is new.
-/python/ray/_common/ @edoakes @aslonnie
+/python/ray/_common/ @edoakes @jjyao
 
 # Ray data.
 /python/ray/data/ @ray-project/ray-data
@@ -67,6 +67,7 @@
 
 # LLM
 /python/ray/llm/ @ray-project/ray-llm
+/python/ray/data/llm.py @ray-project/ray-llm
 
 # Ray Serve
 /python/ray/serve/ @ray-project/ray-serve
@@ -78,8 +79,8 @@
 /python/requirements/ml/dl-gpu-requirements.txt @richardliaw @matthewdeng
 
 # Ray symbol export
-/src/ray/ray_version_script.lds @aslonnie
-/src/ray/ray_exported_symbols.lds @aslonnie
+/src/ray/ray_version_script.lds @ray-project/ray-core
+/src/ray/ray_exported_symbols.lds @ray-project/ray-core
 
 # Ray usage stats
 /python/ray/_private/usage/ @edoakes @richardliaw @jjyao
@@ -111,8 +112,8 @@
 # on their own.
 /release/ray_release/byod/*.sh
 
-/.github/ISSUE_TEMPLATE/ @aslonnie
+/.github/ISSUE_TEMPLATE/ @ray-project/ray-ci
 
 /.github/workflows/ @ray-project/ray-ci
 
-/.gemini/ @edoakes @aslonnie
+/.gemini/ @edoakes @ray-project/ray-ci

.pre-commit-config.yaml

@@ -44,7 +44,6 @@ repos:
         args: [ --fix, --exit-non-zero-on-fix ]
       - id: ruff
         args: [ --select, "I", --fix, --exit-non-zero-on-fix ]
-        files: '^python/ray/serve/|^python/ray/train|^python/ray/data|^python/ray/_private/|^python/ray/llm/|^python/ray/tune/|^python/ray/includes/|^python/ray/internal/|^python/ray/ray_operator/|^python/ray/scripts/|^python/ray/streaming/|^python/ray/dag/|^python/ray/tests/'
 
   - repo: https://github.com/jsh9/pydoclint
     rev: "0.6.6"

.vale/styles/config/vocabularies/General/accept.txt

@@ -33,6 +33,7 @@ autoscales
 bool
 breakpoint
 BTS
+bursty
 chatbot
 CLI
 configs
@@ -45,8 +46,10 @@ deserialize
 deserializes
 dev
 dev to prod
-disable
+[d|D]isable[d]
+[d|D]isable
 DLinear
+Dockerfile
 DPO
 EKS
 ETDataset
@@ -69,13 +72,15 @@ LMs
 LSH
 MCP
 Megatron
+Mixtral
 MLflow
 MLOps
 namespace
 NER
 Nsight
 NumPy
 NVIDIA
+NVLink
 OOM
 open-source
 PACK
@@ -86,6 +91,8 @@ pretraining
 productionize
 Pythonic
 QPS
+Qwen
+Quantizing
 retrigger
 RISECamp
 RLHF
@@ -104,6 +111,7 @@ teardown
 uncaptured
 URI(s)?
 UUID
+USD
 uv
 verl
 VM(s)?

BUILD.bazel

@@ -130,7 +130,7 @@ ray_cc_library(
     ),
     deps = [
         "//src/ray/common:asio",
-        "//src/ray/raylet_client:raylet_client_interface",
+        "//src/ray/rpc:raylet_client_interface",
     ],
 )
 

bazel/gen_extract.py

@@ -1,7 +1,7 @@
-from typing import List, Optional
 import os
 import shutil
 import subprocess
+from typing import List, Optional
 
 import runfiles
 

ci/build/build-manylinux-forge.sh

@@ -5,9 +5,22 @@ set -exuo pipefail
 
 BAZELISK_VERSION="v1.26.0"
 
-platform="linux"
+ARCH="$(uname -m)"
 
-echo "Architecture(HOSTTYPE) is ${HOSTTYPE}"
+case "$ARCH" in
+  x86_64|amd64)
+    ARCH="x86_64"
+    ;;
+  aarch64|arm64)
+    ARCH="aarch64"
+    ;;
+  *)
+    echo "Unsupported arch: $ARCH" >&2
+    exit 1
+    ;;
+esac
+
+echo "Architecture is ${ARCH}"
 
 if [[ ! -e /usr/bin/nproc ]]; then
   echo -e '#!/bin/bash\necho 10' > "/usr/bin/nproc"
@@ -16,7 +29,7 @@ fi
 
 # Install ray cpp dependencies.
 sudo yum -y install unzip zip sudo openssl xz
-if [[ "${HOSTTYPE-}" == "x86_64" ]]; then
+if [[ "${ARCH}" == "x86_64" ]]; then
   sudo yum -y install libasan-4.8.5-44.el7.x86_64 libubsan-7.3.1-5.10.el7.x86_64 \
     devtoolset-8-libasan-devel.x86_64
 fi
@@ -30,28 +43,36 @@ if [[ "${RAYCI_DISABLE_JAVA:-false}" != "true" && "${RAY_INSTALL_JAVA:-1}" == "1
   export JAVA_HOME="${JAVA_BIN%jre/bin/java}"
 fi
 
-# Install ray dashboard dependencies.
-curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.34.0/install.sh | bash
-source "$HOME"/.nvm/nvm.sh
+# Install nodejs
+NODE_VERSION_FULL="${NODE_VERSION_FULL:-14.21.3}"
+
+if [[ "${ARCH}" == "x86_64" ]]; then
+  NODE_URL="https://nodejs.org/dist/v${NODE_VERSION_FULL}/node-v${NODE_VERSION_FULL}-linux-x64.tar.xz"
+  NODE_SHA256="05c08a107c50572ab39ce9e8663a2a2d696b5d262d5bd6f98d84b997ce932d9a"
+else # aarch64
+  NODE_URL="https://nodejs.org/dist/v${NODE_VERSION_FULL}/node-v${NODE_VERSION_FULL}-linux-arm64.tar.xz"
+  NODE_SHA256="f06642bfcf0b8cc50231624629bec58b183954641b638e38ed6f94cd39e8a6ef"
+fi
 
-NODE_VERSION="14"
-nvm install "$NODE_VERSION"
-nvm use "$NODE_VERSION"
+NODE_DIR="/usr/local/node"
+curl -fsSL "${NODE_URL}" -o /tmp/node.tar.xz
+echo "$NODE_SHA256  /tmp/node.tar.xz" | sha256sum -c -
+sudo mkdir -p "$NODE_DIR"
+sudo tar -xf /tmp/node.tar.xz -C "$NODE_DIR" --strip-components=1
+rm /tmp/node.tar.xz
 
 # Install bazel
 mkdir -p "$HOME"/bin
-if [[ "${HOSTTYPE}" == "aarch64" || "${HOSTTYPE}" = "arm64" ]]; then
-  # architecture is "aarch64", but the bazel tag is "arm64"
-  BAZELISK_URL="https://github.com/bazelbuild/bazelisk/releases/download/${BAZELISK_VERSION}/bazelisk-${platform}-arm64"
-elif [[ "${HOSTTYPE}" == "x86_64" ]]; then
-  BAZELISK_URL="https://github.com/bazelbuild/bazelisk/releases/download/${BAZELISK_VERSION}/bazelisk-${platform}-amd64"
-else
-  echo "Could not found matching bazelisk URL for platform ${platform} and architecture ${HOSTTYPE}"
-  exit 1
+if [[ "${ARCH}" == "x86_64" ]]; then
+  BAZELISK_URL="https://github.com/bazelbuild/bazelisk/releases/download/${BAZELISK_VERSION}/bazelisk-linux-amd64"
+else # aarch64
+  BAZELISK_URL="https://github.com/bazelbuild/bazelisk/releases/download/${BAZELISK_VERSION}/bazelisk-linux-arm64"
 fi
-curl -sSfL -o "$HOME"/bin/bazelisk "${BAZELISK_URL}"
-chmod +x "$HOME"/bin/bazelisk
-sudo ln -sf "$HOME"/bin/bazelisk /usr/local/bin/bazel
+
+curl -sSfL -o /tmp/bazelisk "${BAZELISK_URL}"
+chmod +x /tmp/bazelisk
+sudo mv /tmp/bazelisk /usr/local/bin/bazelisk
+sudo ln -sf /usr/local/bin/bazelisk /usr/local/bin/bazel
 
 # Use python3.9 as default python3
 sudo ln -sf /usr/local/bin/python3.9 /usr/local/bin/python3

ci/build/build-manylinux-ray.sh

@@ -11,7 +11,11 @@ if [[ "${RAY_INSTALL_JAVA}" == "1" ]]; then
   bazel build //java:ray_java_pkg
 fi
 
+export PATH="/usr/local/node/bin:$PATH"
+
 # Build ray dashboard
-cd python/ray/dashboard/client
-npm ci
-npm run build
+(
+  cd python/ray/dashboard/client
+  npm ci
+  npm run build
+)