Addresses: Backend session cache not cluster-safe (multi-worker issue) Problem: - Session cache is process-local (InMemorySessionCache) - Multi-worker deployments (uvicorn --workers N) create separate processes - Each process has its own independent session cache - Sessions cached in Worker A are invisible to Workers B, C, D - Users randomly logged out when requests land on different workers - Also affects RuntimeState, rate limiter, and background jobs Solution (Option A - Strict single-worker enforcement): - Enhance startup validation with clearer error messages - Update error messages to explain the problem and how to fix it - Document single-worker requirement prominently in Docker configs - Update module docstrings to clarify constraints Changes: 1. app/startup.py: - Enhanced _check_single_worker_mode() error message with troubleshooting - Enhanced _stage_check_worker_mode_and_acquire_lock() error message - Removed unused import 2. app/utils/session_cache.py: - Updated module docstring to explain constraints more clearly - Added references to deployment documentation - Clarified multi-worker solution for future implementation 3. app/utils/runtime_state.py: - Updated module docstring with deployment constraint references - Aligned messaging with session_cache.py 4. Docker/Dockerfile.backend: - Added comprehensive comments about single-worker requirement - Explained impact in multi-worker deployments - Referenced deployment constraints documentation 5. Docker/docker-compose.yml, compose.prod.yml, compose.debug.yml: - Added documentation comments about BANGUI_WORKERS constraint - Explained why single-worker is required 6. backend/tests/test_startup_integration.py: - Fixed test unpacking to match function return signature (3 values, not 2) This ensures multi-worker deployments fail loudly at startup with clear guidance on what went wrong and how to fix it. The database-backed scheduler lock provides defense-in-depth for container orchestration scenarios. For future multi-worker support, implement: - Redis or database-backed session cache - Shared RuntimeState coordination - Distributed APScheduler backend Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
125 lines
4.5 KiB
Python
125 lines
4.5 KiB
Python
"""Pluggable session cache abstraction.
|
|
|
|
This module defines a cache interface for authenticated sessions and a default
|
|
process-local in-memory implementation. The backend can swap the cache
|
|
implementation without changing the authentication dependency logic.
|
|
|
|
⚠️ PROCESS-LOCAL CONSTRAINT (InMemorySessionCache)
|
|
====================================================
|
|
|
|
InMemorySessionCache stores validated sessions in a process-local dict.
|
|
This means:
|
|
- Each worker process has its own independent session cache.
|
|
- A session invalidated (logout) by worker A is still valid in worker B.
|
|
- Changes to the cache in one process are NOT visible to other processes.
|
|
|
|
IMPACT IN MULTI-WORKER DEPLOYMENTS:
|
|
- User logs out (worker A clears session from its cache).
|
|
- User makes a new request → routed to worker B.
|
|
- Worker B still has the stale session in its cache → request is accepted.
|
|
- User appears still logged in (from their perspective).
|
|
|
|
This is a CRITICAL SECURITY ISSUE: logout does not work reliably across workers.
|
|
|
|
SINGLE-WORKER ENFORCEMENT:
|
|
BanGUI enforces single-worker mode to prevent this issue:
|
|
1. Environment variable check: BANGUI_WORKERS must be 1 or unset
|
|
2. Database lock: Only one instance can run the scheduler at a time
|
|
3. Startup validation: Fails loudly if multi-worker scenario is detected
|
|
|
|
See Docs/Architekture.md § Deployment Constraints for full details.
|
|
|
|
MULTI-WORKER SOLUTION (Future):
|
|
If multi-worker support is needed in the future, replace InMemorySessionCache
|
|
with a shared backend such as:
|
|
- RedisSessionCache — backed by Redis (recommended for production)
|
|
- DatabaseSessionCache — backed by SQLite or PostgreSQL
|
|
- SharedMemorySessionCache — backed by IPC (for local multi-process)
|
|
|
|
The SessionCache Protocol is designed for pluggable backends:
|
|
class SessionCache(Protocol):
|
|
def get(token: str) -> Session | None: ...
|
|
def set(token: str, session: Session, ttl_seconds: float) -> None: ...
|
|
def invalidate(token: str) -> None: ...
|
|
def clear() -> None: ...
|
|
|
|
To add Redis support:
|
|
1. Create RedisSessionCache in this module (implements SessionCache)
|
|
2. Update app/main.py _update_session_cache() to instantiate RedisSessionCache
|
|
when BANGUI_REDIS_URL is configured
|
|
3. Update Backend-Development.md with multi-worker deployment guidelines
|
|
|
|
CURRENT STATUS:
|
|
For now, BanGUI is deployed as single-worker only. This constraint is
|
|
acceptable and keeps the implementation simple. The database-backed scheduler
|
|
lock ensures only one instance runs background jobs, even in container
|
|
orchestration scenarios where multiple instances may start.
|
|
"""
|
|
|
|
from __future__ import annotations
|
|
|
|
import time
|
|
from typing import TYPE_CHECKING, Protocol
|
|
|
|
if TYPE_CHECKING: # pragma: no cover
|
|
from app.models.auth import Session
|
|
|
|
|
|
class SessionCache(Protocol):
|
|
"""Interface for session token validation cache backends."""
|
|
|
|
def get(self, token: str) -> Session | None:
|
|
"""Return the cached session for *token*, or ``None`` if missing."""
|
|
|
|
def set(self, token: str, session: Session, ttl_seconds: float) -> None:
|
|
"""Cache the validated *session* for *token* for *ttl_seconds*."""
|
|
|
|
def invalidate(self, token: str) -> None:
|
|
"""Remove *token* from the cache if it exists."""
|
|
|
|
def clear(self) -> None:
|
|
"""Remove all entries from the cache."""
|
|
|
|
|
|
class InMemorySessionCache:
|
|
"""A process-local session cache implementation."""
|
|
|
|
def __init__(self) -> None:
|
|
self._entries: dict[str, tuple[Session, float]] = {}
|
|
|
|
def get(self, token: str) -> Session | None:
|
|
entry = self._entries.get(token)
|
|
if entry is None:
|
|
return None
|
|
|
|
session, expires_at = entry
|
|
if time.monotonic() >= expires_at:
|
|
self._entries.pop(token, None)
|
|
return None
|
|
return session
|
|
|
|
def set(self, token: str, session: Session, ttl_seconds: float) -> None:
|
|
self._entries[token] = (session, time.monotonic() + ttl_seconds)
|
|
|
|
def invalidate(self, token: str) -> None:
|
|
self._entries.pop(token, None)
|
|
|
|
def clear(self) -> None:
|
|
self._entries.clear()
|
|
|
|
|
|
class NoOpSessionCache:
|
|
"""A no-op session cache used when caching is disabled."""
|
|
|
|
def get(self, token: str) -> Session | None:
|
|
return None
|
|
|
|
def set(self, token: str, session: Session, ttl_seconds: float) -> None:
|
|
return None
|
|
|
|
def invalidate(self, token: str) -> None:
|
|
return None
|
|
|
|
def clear(self) -> None:
|
|
return None
|