Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Pipeline generator plugins #33

Merged
merged 13 commits into from
Sep 25, 2024
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
110 changes: 110 additions & 0 deletions scripts/pipeline_generator/plugin.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,110 @@
from pydantic import BaseModel, Field
from typing import List, Dict, Any, Optional

from .utils import HF_HOME

DOCKER_PLUGIN_NAME = "docker#v5.2.0"
KUBERNETES_PLUGIN_NAME = "kubernetes"


class DockerPluginConfig(BaseModel):
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

could you add some doc / comments, like for everything?

like maybe some link pointers to where the golden spec of these structs are defined?

"""
Configuration for Docker plugin running in a Buildkite step.
The specification is based on:
https://github.com/buildkite-plugins/docker-buildkite-plugin?tab=readme-ov-file#configuration
"""
image: str = ""
always_pull: bool = Field(default=True, alias="always-pull")
propagate_environment: bool = Field(default=True, alias="propagate-environment")
gpus: Optional[str] = "all"
mount_buildkite_agent: Optional[bool] = Field(default=False, alias="mount-buildkite-agent")
command: List[str] = Field(default_factory=list)
environment: List[str] = [
f"HF_HOME={HF_HOME}",
"VLLM_USAGE_SOURCE=ci-test",
"HF_TOKEN",
"BUILDKITE_ANALYTICS_TOKEN"
]
volumes: List[str] = [
"/dev/shm:/dev/shm",
f"{HF_HOME}:{HF_HOME}"
]


class KubernetesPodContainerConfig(BaseModel):
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

we cancelled the plan to move back to k8s stack? why are we adding this?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is for the A100 node on EKS (roblox aws)

"""
Configuration for a container running in a Kubernetes pod.
"""
image: str
command: List[str]
resources: Dict[str, Dict[str, int]]
volume_mounts: List[Dict[str, str]] = Field(
alias="volumeMounts",
default=[
{"name": "devshm", "mountPath": "/dev/shm"},
{"name": "hf-cache", "mountPath": HF_HOME}
]
)
env: List[Dict[str, str]] = Field(
default=[
{"name": "HF_HOME", "value": HF_HOME},
{"name": "VLLM_USAGE_SOURCE", "value": "ci-test"},
{
"name": "HF_TOKEN",
"valueFrom": {
"secretKeyRef": {
"name": "hf-token-secret",
"key": "token"
}
}
},
],
)


class KubernetesPodSpec(BaseModel):
"""
Configuration for a Kubernetes pod running in a Buildkite step.
"""
containers: List[KubernetesPodContainerConfig]
priority_class_name: str = Field(default="ci", alias="priorityClassName")
node_selector: Dict[str, Any] = Field(
default={"nvidia.com/gpu.product": "NVIDIA-A100-SXM4-80GB"},
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

where does this NVIDIA-A100-SXM4-80GB come from?

alias="nodeSelector"
)
volumes: List[Dict[str, Any]] = Field(
default=[
{"name": "devshm", "emptyDir": {"medium": "Memory"}},
{"name": "hf-cache", "hostPath": {"path": HF_HOME, "type": "Directory"}}
]
)


class KubernetesPluginConfig(BaseModel):
"""
Configuration for Kubernetes plugin running in a Buildkite step.
"""
pod_spec: KubernetesPodSpec = Field(alias="podSpec")


def get_kubernetes_plugin_config(container_image: str, test_bash_command: List[str], num_gpus: int) -> Dict:
pod_spec = KubernetesPodSpec(
containers=[
KubernetesPodContainerConfig(
image=container_image,
command=[" ".join(test_bash_command)],
resources={"limits": {"nvidia.com/gpu": num_gpus}}
)
]
)
return {KUBERNETES_PLUGIN_NAME: KubernetesPluginConfig(podSpec=pod_spec).dict(by_alias=True)}


def get_docker_plugin_config(docker_image_path: str, test_bash_command: List[str], no_gpu: bool) -> Dict:
docker_plugin_config = DockerPluginConfig(
image=docker_image_path,
command=test_bash_command
)
if no_gpu:
docker_plugin_config.gpus = None
return {DOCKER_PLUGIN_NAME: docker_plugin_config.dict(exclude_none=True, by_alias=True)}
116 changes: 116 additions & 0 deletions scripts/tests/pipeline_generator/test_plugin.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,116 @@
import pytest
import sys

from scripts.pipeline_generator.plugin import (
get_kubernetes_plugin_config,
get_docker_plugin_config,
DOCKER_PLUGIN_NAME,
KUBERNETES_PLUGIN_NAME,
)


def test_get_kubernetes_plugin_config():
docker_image_path = "test_image:latest"
test_bash_command = ["echo", "Hello, Kubernetes!"]
num_gpus = 1

expected_config = {
KUBERNETES_PLUGIN_NAME: {
"podSpec": {
"containers": [
{
"image": docker_image_path,
"command": [" ".join(test_bash_command)],
"resources": {"limits": {"nvidia.com/gpu": num_gpus}},
"volumeMounts": [
{"name": "devshm", "mountPath": "/dev/shm"},
{"name": "hf-cache", "mountPath": "/root/.cache/huggingface"}
],
"env": [
{"name": "HF_HOME", "value": "/root/.cache/huggingface"},
{"name": "VLLM_USAGE_SOURCE", "value": "ci-test"},
{
"name": "HF_TOKEN",
"valueFrom": {
"secretKeyRef": {
"name": "hf-token-secret",
"key": "token"
}
}
},
],
}
],
"priorityClassName": "ci",
"nodeSelector": {"nvidia.com/gpu.product": "NVIDIA-A100-SXM4-80GB"},
"volumes": [
{"name": "devshm", "emptyDir": {"medium": "Memory"}},
{"name": "hf-cache", "hostPath": {"path": "/root/.cache/huggingface", "type": "Directory"}}
]
}
}
}

assert get_kubernetes_plugin_config(docker_image_path, test_bash_command, num_gpus) == expected_config


@pytest.mark.parametrize(
"docker_image_path, test_bash_command, no_gpu, expected_config",
[
(
"test_image:latest",
["bash", "-c", "echo A;\npytest -v -s a.py"],
False,
{
DOCKER_PLUGIN_NAME: {
"image": "test_image:latest",
"always-pull": True,
"propagate-environment": True,
"gpus": "all",
"command": ["bash", "-c", "echo A;\npytest -v -s a.py"],
"environment": [
"HF_HOME=/root/.cache/huggingface",
"VLLM_USAGE_SOURCE=ci-test",
"HF_TOKEN",
"BUILDKITE_ANALYTICS_TOKEN"
],
"mount-buildkite-agent": False,
"volumes": [
"/dev/shm:/dev/shm",
"/root/.cache/huggingface:/root/.cache/huggingface"
]
}
}
),
(
"cpu_image:latest",
["bash", "-c", "echo B;\npytest -v -s b.py"],
True,
{
DOCKER_PLUGIN_NAME: {
"image": "cpu_image:latest",
"always-pull": True,
"propagate-environment": True,
"command": ["bash", "-c", "echo B;\npytest -v -s b.py"],
"environment": [
"HF_HOME=/root/.cache/huggingface",
"VLLM_USAGE_SOURCE=ci-test",
"HF_TOKEN",
"BUILDKITE_ANALYTICS_TOKEN"
],
"mount-buildkite-agent": False,
"volumes": [
"/dev/shm:/dev/shm",
"/root/.cache/huggingface:/root/.cache/huggingface"
]
}
}
),
]
)
def test_get_docker_plugin_config(docker_image_path, test_bash_command, no_gpu, expected_config):
assert get_docker_plugin_config(docker_image_path, test_bash_command, no_gpu) == expected_config


if __name__ == "__main__":
sys.exit(pytest.main(["-v", __file__]))
Loading