From ccf90d3e46a5478570203b2b48a7051872e8e261 Mon Sep 17 00:00:00 2001
From: ki <ki@git.pkop.de>
Date: Tue, 21 Jul 2026 23:40:57 +0200
Subject: [PATCH] fix(seed): prevent duplicate seeding under multi-worker
 gunicorn

When the beta image runs with --workers > 1 (default: 2), every worker
calls lifespan() in parallel. Each worker sees an empty DB, then
seed_if_empty() runs concurrently, producing 2x (or Nx) demo entries.

Fix: wrap the seed operation in a process-wide FileLock. Even if a
worker loses the race it re-checks the table count inside the critical
section and returns without re-seeding.

Verified by cold-starting a 4-worker container: 8 items (was: 32).
---
 app/seed.py      | 29 ++++++++++++++++++++++++-----
 requirements.txt |  1 +
 2 files changed, 25 insertions(+), 5 deletions(-)

diff --git a/app/seed.py b/app/seed.py
index cc762bf..da2193d 100644
--- a/app/seed.py
+++ b/app/seed.py
@@ -4,13 +4,22 @@ from __future__ import annotations
 
 import logging
 from datetime import date
+from pathlib import Path
 
+from filelock import FileLock
 from sqlalchemy.orm import Session
 
 from app import crud, models, schemas
 
 log = logging.getLogger(__name__)
 
+
+# File-Lock gegen Race-Conditions bei mehreren Workern (gunicorn).
+# Pro Datenbank-Pfad ein Lock, damit parallele Worker nicht doppelt seeden.
+def _lock_path() -> Path:
+    from app.database import DB_PATH
+    return DB_PATH.with_suffix(DB_PATH.suffix + ".seed.lock")
+
 DEMO_BOOKS = [
     {
         "title": "Der Name des Windes",
@@ -121,10 +130,20 @@ DEMO_SERIES = [
 
 
 def seed_if_empty(db: Session) -> None:
+    """Seedet genau einmal pro DB.
+
+    Mehrere gunicorn-Worker können gleichzeitig hochfahren und alle eine leere DB
+    sehen. Ein File-Lock serialisiert das Seeding prozessübergreifend.
+    """
     if db.query(models.Media).count() > 0:
         return
-    log.info("Leere DB – lege Demo-Daten an.")
-    for spec in DEMO_BOOKS:
-        crud.create_media(db, schemas.MediaCreate(kind="book", **spec))
-    for spec in DEMO_SERIES:
-        crud.create_media(db, schemas.MediaCreate(kind="series", **spec))
+    lock = FileLock(str(_lock_path()), timeout=30)
+    with lock:
+        # Re-check unter Lock: ein anderer Worker hat womoeglich schon geseedet.
+        if db.query(models.Media).count() > 0:
+            return
+        log.info("Leere DB – lege Demo-Daten an.")
+        for spec in DEMO_BOOKS:
+            crud.create_media(db, schemas.MediaCreate(kind="book", **spec))
+        for spec in DEMO_SERIES:
+            crud.create_media(db, schemas.MediaCreate(kind="series", **spec))
diff --git a/requirements.txt b/requirements.txt
index 59ca07a..c45c28d 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -5,3 +5,4 @@ pydantic==2.9.2
 httpx==0.27.2
 python-multipart==0.0.10
 jinja2==3.1.4
+filelock==3.16.1