[Bugfix] Ensure stage and diffusion subprocesses exit when parent dies unexpectedly (#3751)

Signed-off-by: Lancer <maruixiang6688@gmail.com>
Co-authored-by: Alicia <115451386+congw729@users.noreply.github.com>
This commit is contained in:
LancerandAlicia authored and GitHub committed 2026-05-28 08:22:04 +08:00
1 parent 1f3a3e3848
commit 5c3c81920b
12 files changed
+270 -36

No files matched your search

@@ -11,6 +11,7 @@ to DiffusionModelRunner.
import gc
import multiprocessing as mp
import os
import signal
from collections.abc import Iterable, Iterator
from contextlib import AbstractContextManager, contextmanager, nullcontext
from dataclasses import dataclass
@@ -46,6 +47,7 @@ from vllm_omni.diffusion.request import OmniDiffusionRequest
from vllm_omni.diffusion.sched.interface import DiffusionSchedulerOutput
from vllm_omni.diffusion.worker.diffusion_model_runner import DiffusionModelRunner
from vllm_omni.diffusion.worker.utils import BaseRunnerOutput
from vllm_omni.engine.stage_init_utils import set_death_signal
from vllm_omni.lora.request import LoRARequest
from vllm_omni.platforms import current_omni_platform
from vllm_omni.profiler import OmniTorchProfilerWrapper, create_omni_profiler
@@ -831,6 +833,8 @@ class WorkerProc:
"""Worker initialization and execution loops."""
from vllm_omni.plugins import load_omni_general_plugins
set_death_signal(signal.SIGTERM)
# Set process title for visibility in nvidia-smi / htop (optional, non-fatal)
try:
import setproctitle