e489ad7a21
- **Add SPDX license headers to python source files** - **Check for SPDX headers using pre-commit** commit 9d7ef44c3cfb72ca4c32e1c677d99259d10d4745 Author: Russell Bryant <rbryant@redhat.com> Date: Fri Jan 31 14:18:24 2025 -0500 Add SPDX license headers to python source files This commit adds SPDX license headers to python source files as recommended to the project by the Linux Foundation. These headers provide a concise way that is both human and machine readable for communicating license information for each source file. It helps avoid any ambiguity about the license of the code and can also be easily used by tools to help manage license compliance. The Linux Foundation runs license scans against the codebase to help ensure we are in compliance with the licenses of the code we use, including dependencies. Having these headers in place helps that tool do its job. More information can be found on the SPDX site: - https://spdx.dev/learn/handling-license-info/ Signed-off-by: Russell Bryant <rbryant@redhat.com> commit 5a1cf1cb3b80759131c73f6a9dddebccac039dea Author: Russell Bryant <rbryant@redhat.com> Date: Fri Jan 31 14:36:32 2025 -0500 Check for SPDX headers using pre-commit Signed-off-by: Russell Bryant <rbryant@redhat.com> --------- Signed-off-by: Russell Bryant <rbryant@redhat.com>
106 lines
3.8 KiB
Python
106 lines
3.8 KiB
Python
# SPDX-License-Identifier: Apache-2.0
|
|
|
|
import asyncio
|
|
import signal
|
|
from http import HTTPStatus
|
|
from typing import Any
|
|
|
|
import uvicorn
|
|
from fastapi import FastAPI, Request, Response
|
|
|
|
from vllm import envs
|
|
from vllm.engine.async_llm_engine import AsyncEngineDeadError
|
|
from vllm.engine.multiprocessing import MQEngineDeadError
|
|
from vllm.logger import init_logger
|
|
from vllm.utils import find_process_using_port
|
|
|
|
logger = init_logger(__name__)
|
|
|
|
|
|
async def serve_http(app: FastAPI, **uvicorn_kwargs: Any):
|
|
logger.info("Available routes are:")
|
|
for route in app.routes:
|
|
methods = getattr(route, "methods", None)
|
|
path = getattr(route, "path", None)
|
|
|
|
if methods is None or path is None:
|
|
continue
|
|
|
|
logger.info("Route: %s, Methods: %s", path, ', '.join(methods))
|
|
|
|
config = uvicorn.Config(app, **uvicorn_kwargs)
|
|
server = uvicorn.Server(config)
|
|
_add_shutdown_handlers(app, server)
|
|
|
|
loop = asyncio.get_running_loop()
|
|
|
|
server_task = loop.create_task(server.serve())
|
|
|
|
def signal_handler() -> None:
|
|
# prevents the uvicorn signal handler to exit early
|
|
server_task.cancel()
|
|
|
|
async def dummy_shutdown() -> None:
|
|
pass
|
|
|
|
loop.add_signal_handler(signal.SIGINT, signal_handler)
|
|
loop.add_signal_handler(signal.SIGTERM, signal_handler)
|
|
|
|
try:
|
|
await server_task
|
|
return dummy_shutdown()
|
|
except asyncio.CancelledError:
|
|
port = uvicorn_kwargs["port"]
|
|
process = find_process_using_port(port)
|
|
if process is not None:
|
|
logger.debug(
|
|
"port %s is used by process %s launched with command:\n%s",
|
|
port, process, " ".join(process.cmdline()))
|
|
logger.info("Shutting down FastAPI HTTP server.")
|
|
return server.shutdown()
|
|
|
|
|
|
def _add_shutdown_handlers(app: FastAPI, server: uvicorn.Server) -> None:
|
|
"""Adds handlers for fatal errors that should crash the server"""
|
|
|
|
@app.exception_handler(RuntimeError)
|
|
async def runtime_error_handler(request: Request, __):
|
|
"""On generic runtime error, check to see if the engine has died.
|
|
It probably has, in which case the server will no longer be able to
|
|
handle requests. Trigger a graceful shutdown with a SIGTERM."""
|
|
engine = request.app.state.engine_client
|
|
if (not envs.VLLM_KEEP_ALIVE_ON_ENGINE_DEATH and engine.errored
|
|
and not engine.is_running):
|
|
logger.fatal("AsyncLLMEngine has failed, terminating server "
|
|
"process")
|
|
# See discussions here on shutting down a uvicorn server
|
|
# https://github.com/encode/uvicorn/discussions/1103
|
|
# In this case we cannot await the server shutdown here because
|
|
# this handler must first return to close the connection for
|
|
# this request.
|
|
server.should_exit = True
|
|
|
|
return Response(status_code=HTTPStatus.INTERNAL_SERVER_ERROR)
|
|
|
|
@app.exception_handler(AsyncEngineDeadError)
|
|
async def async_engine_dead_handler(_, __):
|
|
"""Kill the server if the async engine is already dead. It will
|
|
not handle any further requests."""
|
|
if not envs.VLLM_KEEP_ALIVE_ON_ENGINE_DEATH:
|
|
logger.fatal("AsyncLLMEngine is already dead, terminating server "
|
|
"process")
|
|
server.should_exit = True
|
|
|
|
return Response(status_code=HTTPStatus.INTERNAL_SERVER_ERROR)
|
|
|
|
@app.exception_handler(MQEngineDeadError)
|
|
async def mq_engine_dead_handler(_, __):
|
|
"""Kill the server if the mq engine is already dead. It will
|
|
not handle any further requests."""
|
|
if not envs.VLLM_KEEP_ALIVE_ON_ENGINE_DEATH:
|
|
logger.fatal("MQLLMEngine is already dead, terminating server "
|
|
"process")
|
|
server.should_exit = True
|
|
|
|
return Response(status_code=HTTPStatus.INTERNAL_SERVER_ERROR)
|