Question

我在Docker中运行Python Tornado应用程序，而部分API涉及连接到DocumentDB以进行存储：

client = document_client.DocumentClient(config.uri, {'masterKey': config.key})
db = next((data for data in client.ReadDatabases() if data['id'] == config.db))
coll = next((docs for docs in client.ReadCollections(db['_self']) if docs['id'] == config.collection))

授权工作正常，我通过添加和删除文档完成了对数据库的多次调用。当我离开Docker容器运行几个小时（没有准确计算需要多长时间）或者当我将容器放置过夜并在早上检查它时，问题出现了，我得到了这个错误：

Traceback (most recent call last):
tornado1_1  |   File "api_app.py", line 76, in <module>
tornado1_1  |     class UserHandler(BaseHandler):
tornado1_1  |   File "api_app.py", line 82, in UserHandler
tornado1_1  |     db = next((data for data in client.ReadDatabases() if data['id'] == config.db))
tornado1_1  |   File "api_app.py", line 82, in <genexpr>
tornado1_1  |     db = next((data for data in client.ReadDatabases() if data['id'] == config.db))
tornado1_1  |   File "/usr/local/lib/python2.7/site-packages/pydocumentdb/query_iterable.py", line 123, in next
tornado1_1  |     retry_utility._Execute(self._iterable._client, self._iterable._client._global_endpoint_manager, callback)
tornado1_1  |   File "/usr/local/lib/python2.7/site-packages/pydocumentdb/retry_utility.py", line 48, in _Execute
tornado1_1  |     result = _ExecuteFunction(function, *args, **kwargs)
tornado1_1  |   File "/usr/local/lib/python2.7/site-packages/pydocumentdb/retry_utility.py", line 81, in _ExecuteFunction
tornado1_1  |     return function(*args, **kwargs)
tornado1_1  |   File "/usr/local/lib/python2.7/site-packages/pydocumentdb/query_iterable.py", line 114, in callback
tornado1_1  |     if not self._iterable.fetch_next_block():
tornado1_1  |   File "/usr/local/lib/python2.7/site-packages/pydocumentdb/query_iterable.py", line 144, in fetch_next_block
tornado1_1  |     fetched_items = self.fetch_items()
tornado1_1  |   File "/usr/local/lib/python2.7/site-packages/pydocumentdb/query_iterable.py", line 184, in fetch_items
tornado1_1  |     (fetched_items, response_headers) = self._fetch_function(self._options)
tornado1_1  |   File "/usr/local/lib/python2.7/site-packages/pydocumentdb/document_client.py", line 225, in fetch_fn
tornado1_1  |     options), self.last_response_headers
tornado1_1  |   File "/usr/local/lib/python2.7/site-packages/pydocumentdb/document_client.py", line 2349, in __QueryFeed
tornado1_1  |     headers)
tornado1_1  |   File "/usr/local/lib/python2.7/site-packages/pydocumentdb/document_client.py", line 2206, in __Get
tornado1_1  |     headers)
tornado1_1  |   File "/usr/local/lib/python2.7/site-packages/pydocumentdb/synchronized_request.py", line 168, in SynchronizedRequest
tornado1_1  |     return retry_utility._Execute(client, global_endpoint_manager, _InternalRequest, connection_policy, request_options, request_body)
tornado1_1  |   File "/usr/local/lib/python2.7/site-packages/pydocumentdb/retry_utility.py", line 48, in _Execute
tornado1_1  |     result = _ExecuteFunction(function, *args, **kwargs)
tornado1_1  |   File "/usr/local/lib/python2.7/site-packages/pydocumentdb/retry_utility.py", line 81, in _ExecuteFunction
tornado1_1  |     return function(*args, **kwargs)
tornado1_1  |   File "/usr/local/lib/python2.7/site-packages/pydocumentdb/synchronized_request.py", line 100, in _InternalRequest
tornado1_1  |     raise errors.HTTPFailure(response.status, data, headers)
tornado1_1  | pydocumentdb.errors.HTTPFailure: Status code: 403
tornado1_1  | {"code":"Forbidden","message":"The authorization token is not valid at the current time. Please create another token and retry (token start time: Thu, 04 Aug 2016 04:30:53 GMT, token expiry time: Thu, 04 Aug 2016 04:45:53 GMT, current server time: Thu, 04 Aug 2016 15:11:11 GMT).\r\nActivityId: af4c602a-9413-4eb3-b270-b8a57fa2d973"}

正如您所看到的，它可以与客户端建立连接，但它在db = next((data for data in client.ReadDatabases() if data['id'] == config.db))行失败，并引发一些关于服务器与令牌开始时间之间不匹配的奇怪错误。一旦我重新启动计算机（不仅仅是容器），它将再次运行一段不确定的时间。我在Azure Documentation上阅读了以下提示：

提示：资源令牌的默认有效时间跨度为1小时。但是，令牌生存期可以明确指定，最长可达5小时。

不确定是否与此有关。

Answer 1

这可能是由于您的机器时间漂移（与服务器相比）持续增加，直到差异为一小时。

在异常消息中，您可以看到令牌开始/结束时间与当前服务器时间之间的延迟。

Answer 2

这听起来像授权令牌到期。因此，您需要生成另一个。错误消息显示“请创建另一个令牌并重试”。

也许你在创建容器时创建了令牌？您可以尝试删除容器以强制它创建一个新容器。

Answer 3

根据HTTP Status Codes for DocumentDB和您的说明，异常信息的状态代码403表示 The authorization token expired.

所以解决方案就是通过catch＆amp; amp;创建一个新的客户端连接，而不是旧的客户端连接。处理异常。

作为参考，文章末尾有一个提示。

提示

资源令牌的默认有效时间跨度为1小时。但是，令牌生存期可以明确指定，最长可达5小时。

您可以通过指定标头x-ms-documentdb-expiry-seconds的值来引用REST API Create a Permission或Replace a Permission来修改令牌生存期。

Docker中的DocumentDB - ＆＃34;授权令牌在当前时间无效。＆＃34;

3 个答案: