feat: implement graceful shutdown with SIGINT/SIGTERM support
- Add WebSocket shutdown() with client notification and graceful close - Enhance download service stop() with pending state persistence - Expand FastAPI lifespan shutdown with proper cleanup sequence - Add SQLite WAL checkpoint before database close - Update stop_server.sh to use SIGTERM with timeout fallback - Configure uvicorn timeout_graceful_shutdown=30s - Update ARCHITECTURE.md with shutdown documentation
This commit is contained in:
@@ -997,30 +997,76 @@ class DownloadService:
|
||||
"""
|
||||
logger.info("Download queue service initialized")
|
||||
|
||||
async def stop(self) -> None:
|
||||
"""Stop the download queue service and cancel active downloads.
|
||||
async def stop(self, timeout: float = 10.0) -> None:
|
||||
"""Stop the download queue service gracefully.
|
||||
|
||||
Cancels any active download and shuts down the thread pool immediately.
|
||||
Persists in-progress downloads back to pending state, cancels active
|
||||
tasks, and shuts down the thread pool with a timeout.
|
||||
|
||||
Args:
|
||||
timeout: Maximum time (seconds) to wait for executor shutdown
|
||||
"""
|
||||
logger.info("Stopping download queue service...")
|
||||
logger.info("Stopping download queue service (timeout=%.1fs)...", timeout)
|
||||
|
||||
# Set shutdown flag
|
||||
# Set shutdown flag first to prevent new downloads
|
||||
self._is_shutting_down = True
|
||||
self._is_stopped = True
|
||||
|
||||
# Persist active download back to pending state if one exists
|
||||
if self._active_download:
|
||||
logger.info(
|
||||
"Persisting active download to pending: item_id=%s",
|
||||
self._active_download.id
|
||||
)
|
||||
try:
|
||||
# Reset status to pending so it can be resumed on restart
|
||||
self._active_download.status = DownloadStatus.PENDING
|
||||
self._active_download.completed_at = None
|
||||
await self._save_to_database(self._active_download)
|
||||
logger.info("Active download persisted to database as pending")
|
||||
except Exception as e:
|
||||
logger.error("Failed to persist active download: %s", e)
|
||||
|
||||
# Cancel active download task if running
|
||||
active_task = self._active_download_task
|
||||
if active_task and not active_task.done():
|
||||
logger.info("Cancelling active download task...")
|
||||
active_task.cancel()
|
||||
try:
|
||||
await active_task
|
||||
# Wait briefly for cancellation to complete
|
||||
await asyncio.wait_for(
|
||||
asyncio.shield(active_task),
|
||||
timeout=2.0
|
||||
)
|
||||
except asyncio.TimeoutError:
|
||||
logger.warning("Download task cancellation timed out")
|
||||
except asyncio.CancelledError:
|
||||
logger.info("Active download task cancelled")
|
||||
except Exception as e:
|
||||
logger.warning("Error during task cancellation: %s", e)
|
||||
|
||||
# Shutdown executor immediately, don't wait for tasks
|
||||
# Shutdown executor with wait and timeout
|
||||
logger.info("Shutting down thread pool executor...")
|
||||
self._executor.shutdown(wait=False, cancel_futures=True)
|
||||
try:
|
||||
# Run executor shutdown in thread to avoid blocking event loop
|
||||
loop = asyncio.get_event_loop()
|
||||
await asyncio.wait_for(
|
||||
loop.run_in_executor(
|
||||
None,
|
||||
lambda: self._executor.shutdown(wait=True, cancel_futures=True)
|
||||
),
|
||||
timeout=timeout
|
||||
)
|
||||
logger.info("Thread pool executor shutdown complete")
|
||||
except asyncio.TimeoutError:
|
||||
logger.warning(
|
||||
"Executor shutdown timed out after %.1fs, forcing shutdown",
|
||||
timeout
|
||||
)
|
||||
# Force shutdown without waiting
|
||||
self._executor.shutdown(wait=False, cancel_futures=True)
|
||||
except Exception as e:
|
||||
logger.error("Error during executor shutdown: %s", e)
|
||||
|
||||
logger.info("Download queue service stopped")
|
||||
|
||||
|
||||
Reference in New Issue
Block a user