使用LlamaIndex向Pinecone DB导入数据时遇SSL重试超限错误求助

错误现象 #

URL请求重试次数超限：

(Caused by SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:2427)')))

详细错误栈 #

LlamaIndex_Doc_Helper-JJYEcwwZ\Lib\site-packages\urllib3\util\retry.py", line 515, in increment
    raise MaxRetryError(_pool, url, reason) from reason  # type: ignore[arg-type]
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
urllib3.exceptions.MaxRetryError: HTTPSConnectionPool(host='llamaindex-documentation-helper-ksad0bm.svc.gcp-starter.pinecone.io', port=443): Max retries exceeded with url: /vectors/upsert (Caused by SSLError(SSLEOFError(8, 'EOF occurred in violation of protocol (_ssl.c:2427)')))
Upserted vectors:   0%|          | 0/18 [00:02<?, ?it/s]

使用的Python代码 #

#--- Indexing in Pinecone
# Init. Pinecone Index object
index_name = "llamaindex-documentation-helper"
pinecone_index = pinecone.Index(index_name=index_name)
# Init. vector store object
vector_store = PineconeVectorStore(pinecone_index=pinecone_index)
# Init storage context
storage_context = StorageContext.from_defaults(vector_store=vector_store)
# Creates
index = VectorStoreIndex.from_documents(
    documents=documents, # documents = dir_reader.load_data()
    storage_context=storage_context, # storage_context = StorageContext.from_defaults(vector_store=vector_store)
    service_context=service_context, # service_context = ServiceContext.from_defaults(llm=llm, embed_model=embed_model, node_parser=node_parser)
    show_progress=True,
)

已完成的排查步骤 #

• 关闭防火墙、VPN和杀毒软件
• 系统：Windows 11 Pro
• Python环境：3.12版本（pipenv管理）
• 已升级certifi至最新版本
• 可正常ping通服务器：

LlamaIndex Tutorial\LlamaIndex Doc Helper> ping llamaindex-documentation-helper-ksad0bm.svc.gcp-starter.pinecone.io

Pinging ingress.gcp-starter.pinecone.io [34.160.88.44] with 32 bytes of data:
Reply from 34.160.88.44: bytes=32 time=10ms TTL=114
Reply from 34.160.88.44: bytes=32 time=9ms TTL=114
Reply from 34.160.88.44: bytes=32 time=11ms TTL=114
Reply from 34.160.88.44: bytes=32 time=9ms TTL=114

Ping statistics for 34.160.88.44:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 9ms, Maximum = 11ms, Average = 9ms

解决建议 #

• 降级Python版本：Python 3.12对SSL库的改动可能引发兼容性问题，切换到Python 3.11版本，重新构建pipenv环境后测试。
• 强制指定SSL协议版本：通过自定义HTTP客户端，强制使用TLS 1.2或1.3协议连接Pinecone。示例代码：

  import pinecone
  import requests
  from requests.adapters import HTTPAdapter
  from urllib3.poolmanager import PoolManager
  import ssl
  
  class TLSAdapter(HTTPAdapter):
      def init_poolmanager(self, connections, maxsize, block=False):
          self.poolmanager = PoolManager(
              num_pools=connections,
              maxsize=maxsize,
              block=block,
              ssl_version=ssl.PROTOCOL_TLSv1_3
          )
  
  session = requests.Session()
  session.mount("https://", TLSAdapter())
  
  pinecone.init(api_key="YOUR_API_KEY", environment="YOUR_ENV", session=session)
  pinecone_index = pinecone.Index(index_name=index_name)
  

• 升级Pinecone客户端：执行pip install --upgrade pinecone-client，将客户端更新至最新稳定版。
• 调整超时与重试策略：初始化Pinecone Index时延长超时时间、调整重试次数：

  pinecone_index = pinecone.Index(
      index_name=index_name,
      timeout=60,  # 延长超时时间至60秒
      retry_config={"max_retries": 5}  # 调整最大重试次数
  )
  

• 手动指定证书路径：在代码开头添加以下内容，强制使用certifi提供的证书：

  import os
  import certifi
  os.environ['SSL_CERT_FILE'] = certifi.where()
  

内容的提问来源于stack exchange，提问作者Alex Ricciardi