diff --git a/api/app/celery_worker.py b/api/app/celery_worker.py
index 4ea4fee1..e6c041f5 100644
--- a/api/app/celery_worker.py
+++ b/api/app/celery_worker.py
@@ -2,6 +2,8 @@
 Celery Worker 入口点
 用于启动 Celery Worker: celery -A app.celery_worker worker --loglevel=info
 """
+from celery.signals import worker_process_init
+
 from app.celery_app import celery_app
 from app.core.logging_config import LoggingConfig, get_logger
 
@@ -13,4 +15,19 @@ logger.info("Celery worker logging initialized")
 # 导入任务模块以注册任务
 import app.tasks
 
+
+@worker_process_init.connect
+def _reinit_db_pool(**kwargs):
+    """
+    prefork 子进程启动时重建 SQLAlchemy 连接池。
+    
+    fork() 后子进程继承了父进程的连接池和底层 TCP socket，
+    多个子进程共享同一个 socket 会导致 PostgreSQL 连接损坏。
+    dispose() 会关闭继承来的连接，后续操作会自动创建新连接。
+    """
+    from app.db import engine
+    engine.dispose()
+    logger.info("DB connection pool disposed for forked worker process")
+
+
 __all__ = ['celery_app']
diff --git a/api/app/tasks.py b/api/app/tasks.py
index 8bbbdc6e..87bcaec9 100644
--- a/api/app/tasks.py
+++ b/api/app/tasks.py
@@ -251,8 +251,18 @@ def parse_document(file_path: str, document_id: uuid.UUID):
         # Prepare vision_model for parsing
         vision_model = _build_vision_model(file_path, db_knowledge)
 
+        # 先将文件读入内存，避免解析过程中依赖 NFS 文件持续可访问
+        # python-docx 等库在 binary=None 时会用路径直接打开文件，
+        # 在 NFS/共享存储上可能因缓存失效导致 "Package not found"
+        try:
+            with open(file_path, "rb") as f:
+                file_binary = f.read()
+        except FileNotFoundError:
+            raise FileNotFoundError(f"File not found at '{file_path}'")
+
         from app.core.rag.app.naive import chunk
         res = chunk(filename=file_path,
+                    binary=file_binary,
                     from_page=0,
                     to_page=DEFAULT_PARSE_TO_PAGE,
                     callback=progress_callback,