fix: clean stale partial database backups

.env.example

@@ -177,6 +177,7 @@ EMBEDDING_HOST=
 EMBEDDING_TIMEOUT=30
 OLLAMA_EMBED_MAX_TIMEOUT=30
 OLLAMA_EMBED_KEEP_ALIVE=1m
+PARTIAL_BACKUP_MIN_AGE_MINUTES=60
 OLLAMA_EMBED_MAX_CHARS=4000
 OLLAMA_EMBED_GCP_FAILURE_COOLDOWN_SEC=60
 OLLAMA_EMBED_GCP_FAILURE_NOTICE_SEC=30

config.py

@@ -402,7 +402,7 @@ YOUTUBE_API_KEY = os.getenv('YOUTUBE_API_KEY', '')
 # ==========================================
 # 系統版本與路徑
 # ==========================================
-SYSTEM_VERSION = "V10.627"
+SYSTEM_VERSION = "V10.628"
 LOG_FILE_PATH = os.path.join(BASE_DIR, 'logs/system.log')
 public_url = PUBLIC_URL  # 用於模板顯示
 

docs/memory/current_execution_queue_20260524.md

@@ -330,3 +330,9 @@
 - V10.626 已能在 GCP-A direct timeout 後走 110 proxy，但 cache refresh 仍會先等一次 direct `/api/version` timeout。
 - V10.627 新增 direct-only host health skip：`resolve_ollama_host()` 會讀最近 `host_health_probes`，若 GCP-A/GCP-B direct 在視窗內已 unhealthy，先跳過 direct endpoint，改試同順位 110 proxy；proxy rescue 不吃這個 skip，避免因 direct unhealthy 誤跳過可用 proxy。
 - 新增 `OLLAMA_RESOLVE_HOST_HEALTH_SKIP_ENABLED=true` 與 `OLLAMA_RESOLVE_HOST_HEALTH_SKIP_WINDOW_MINUTES=20`；DB 讀取失敗 fail-open，回到原本網路探測。
+
+## 31. 2026-06-18 V10.628 備份 partial 檔案清理
+
+- 正式 `backup_log` 最新狀態已是 2026-06-18 02:00 成功備份，6/15 的 `pg_dump` not found 是舊失敗紀錄；`get_latest_backup_info()` 目前回 success，backup monitor 不會再因舊 row 告警。
+- 備份目錄仍殘留 0 byte `momo_analytics_*.sql.gz` partial 檔，容易讓人工查檔誤判。
+- V10.628 新增 `cleanup_partial_backups()`，`cleanup_old_backups()` 會先清除超過 `PARTIAL_BACKUP_MIN_AGE_MINUTES=60` 的 0 byte partial 備份；剛產生的 0 byte 檔不刪，避免誤傷正在寫入的備份。

services/db_backup_service.py

@@ -20,6 +20,7 @@ DB_USER = os.environ.get("POSTGRES_USER", "momo")
 DB_NAME = os.environ.get("POSTGRES_DB", "momo_analytics")
 # 保留天數
 RETENTION_DAYS = int(os.environ.get("BACKUP_RETENTION_DAYS", "7"))
+PARTIAL_BACKUP_MIN_AGE_MINUTES = int(os.environ.get("PARTIAL_BACKUP_MIN_AGE_MINUTES", "60"))
 
 
 def _ensure_backup_dir():
@@ -35,6 +36,30 @@ def _remove_partial_backup(filepath: str):
         logger.warning(f"[Backup] 移除不完整備份檔失敗 {filepath}: {exc}")
 
 
+def cleanup_partial_backups(min_age_minutes: int = PARTIAL_BACKUP_MIN_AGE_MINUTES) -> int:
+    """清除失敗後殘留的 0 byte 備份檔，避免監控與人工查檔誤判。"""
+    _ensure_backup_dir()
+    try:
+        min_age_minutes = max(0, int(min_age_minutes))
+    except (TypeError, ValueError):
+        min_age_minutes = 60
+    cutoff = datetime.now() - timedelta(minutes=min_age_minutes)
+    deleted = 0
+    for filepath in glob.glob(os.path.join(BACKUP_DIR, "momo_analytics_*.sql.gz")):
+        try:
+            if os.path.getsize(filepath) != 0:
+                continue
+            mtime = datetime.fromtimestamp(os.path.getmtime(filepath))
+            if mtime > cutoff:
+                continue
+            os.remove(filepath)
+            deleted += 1
+            logger.info(f"[Backup] 已清除 0 byte 不完整備份: {os.path.basename(filepath)}")
+        except Exception as exc:
+            logger.warning(f"[Backup] 清除不完整備份失敗 {filepath}: {exc}")
+    return deleted
+
+
 def _ensure_pg_dump_available() -> str:
     pg_dump_path = shutil.which("pg_dump")
     if pg_dump_path:
@@ -175,10 +200,10 @@ def run_backup() -> dict:
 
 
 def cleanup_old_backups() -> int:
-    """刪除超過保留期限的備份檔，回傳刪除數量"""
+    """刪除不完整與超過保留期限的備份檔，回傳刪除數量。"""
     _ensure_backup_dir()
     cutoff = datetime.now() - timedelta(days=RETENTION_DAYS)
-    deleted = 0
+    deleted = cleanup_partial_backups()
     for f in glob.glob(os.path.join(BACKUP_DIR, "momo_analytics_*.sql.gz")):
         try:
             mtime = datetime.fromtimestamp(os.path.getmtime(f))

tests/test_db_backup_service.py

@@ -0,0 +1,41 @@
+import os
+import time
+
+
+def test_cleanup_partial_backups_removes_only_stale_zero_byte_files(tmp_path, monkeypatch):
+    from services import db_backup_service as backup
+
+    stale_zero = tmp_path / "momo_analytics_20260611_010050.sql.gz"
+    fresh_zero = tmp_path / "momo_analytics_20260618_134900.sql.gz"
+    normal_backup = tmp_path / "momo_analytics_20260618_020001.sql.gz"
+
+    stale_zero.write_bytes(b"")
+    fresh_zero.write_bytes(b"")
+    normal_backup.write_bytes(b"ok")
+
+    old_ts = time.time() - 7200
+    os.utime(stale_zero, (old_ts, old_ts))
+
+    monkeypatch.setattr(backup, "BACKUP_DIR", str(tmp_path))
+
+    deleted = backup.cleanup_partial_backups(min_age_minutes=60)
+
+    assert deleted == 1
+    assert not stale_zero.exists()
+    assert fresh_zero.exists()
+    assert normal_backup.exists()
+
+
+def test_cleanup_old_backups_counts_partial_cleanup(tmp_path, monkeypatch):
+    from services import db_backup_service as backup
+
+    stale_zero = tmp_path / "momo_analytics_20260611_010050.sql.gz"
+    stale_zero.write_bytes(b"")
+    old_ts = time.time() - 7200
+    os.utime(stale_zero, (old_ts, old_ts))
+
+    monkeypatch.setattr(backup, "BACKUP_DIR", str(tmp_path))
+    monkeypatch.setattr(backup, "RETENTION_DAYS", 7)
+
+    assert backup.cleanup_old_backups() == 1
+    assert not stale_zero.exists()