修復 AI embedding queue 原子取件

services/openclaw_learning_service.py

@@ -174,42 +174,65 @@ def _process_one_embedding(row_id: int, target_table: str, target_id: int,
         session.close()
 
 
+def _claim_pending_embeddings(limit: int = EMBED_BATCH_SIZE,
+                              max_attempts: int = EMBED_MAX_ATTEMPTS):
+    """Atomically claim pending embedding jobs across multiple workers."""
+    session = get_session()
+    try:
+        session.execute(
+            text("""
+                UPDATE embedding_retry_queue
+                   SET status = 'pending',
+                       updated_at = :now,
+                       last_error = COALESCE(last_error, '') || ' | reset stale processing'
+                 WHERE status = 'processing'
+                   AND updated_at < :cutoff
+            """),
+            {
+                "now": datetime.now(),
+                "cutoff": datetime.fromtimestamp(time.time() - 15 * 60),
+            },
+        )
+        session.commit()
+        rows = session.execute(
+            text("""
+                WITH picked AS (
+                    SELECT id
+                      FROM embedding_retry_queue
+                     WHERE status = 'pending'
+                       AND attempts < :max
+                     ORDER BY created_at
+                     FOR UPDATE SKIP LOCKED
+                     LIMIT :lim
+                )
+                UPDATE embedding_retry_queue q
+                   SET status = 'processing',
+                       updated_at = :now
+                  FROM picked
+                 WHERE q.id = picked.id
+             RETURNING q.id, q.target_table, q.target_id, q.text_content, q.model
+            """),
+            {
+                "now": datetime.now(),
+                "max": max_attempts,
+                "lim": limit,
+            },
+        ).fetchall()
+        session.commit()
+        return rows
+    except Exception:
+        session.rollback()
+        raise
+    finally:
+        session.close()
+
+
 def _embedding_worker_loop():
     """背景執行緒：輪詢 embedding_retry_queue，批次處理 pending 項目"""
     sys_log.info("[OCLearn] Hermes Embedding Worker (DB-backed) 啟動")
     while True:
         try:
-            session = get_session()
-            try:
-                session.execute(
-                    text("""
-                        UPDATE embedding_retry_queue
-                           SET status = 'pending',
-                               updated_at = :now,
-                               last_error = COALESCE(last_error, '') || ' | reset stale processing'
-                         WHERE status = 'processing'
-                           AND updated_at < :cutoff
-                    """),
-                    {
-                        "now": datetime.now(),
-                        "cutoff": datetime.fromtimestamp(time.time() - 15 * 60),
-                    },
-                )
-                session.commit()
-                rows = session.execute(
-                    text("""
-                        SELECT id, target_table, target_id, text_content, model
-                          FROM embedding_retry_queue
-                         WHERE status = 'pending'
-                           AND attempts < :max
-                         ORDER BY created_at
-                         LIMIT :lim
-                    """),
-                    {"max": EMBED_MAX_ATTEMPTS, "lim": EMBED_BATCH_SIZE},
-                ).fetchall()
-            finally:
-                session.close()
-
+            rows = _claim_pending_embeddings()
             for row in rows:
                 _process_one_embedding(
                     row[0], row[1], row[2], row[3], row[4] or "bge-m3:latest"