1

Dear MongoDB experts,

I’ve been using local deployment for testing and am occasionally seeing the following error in my tests:
````````E pymongo.errors.OperationFailure: Error connecting to Search Index Management service., full error: {‘ok’: 0.0, ‘errmsg’: ‘Error connecting to Search Index Management service.’, ‘code’: 125, ‘codeName’: ‘CommandFailed’, ‘$clusterTime’: {‘clusterTime’: Timestamp(1709934526, 11), ‘signature’: {‘hash’: b’\x13\xda\xce\x9c&>e\xc8\xcf*%\xee\xc6T[\x142\x99\xc0:', ‘keyId’: 7344112858881327110}}, ‘operationTime’: Timestamp(1709934526, 11)}`

This happens when trying to setup a search index on a newly initialized database in local deployment:

python3.11/site-packages/pymongo/collection.py:2435: in create_search_index
    return self.create_search_indexes([model], session, comment, **kwargs)[0]
python3.11/site-packages/pymongo/collection.py:2476: in create_search_indexes
    resp = self._command(
python3.11/site-packages/pymongo/collection.py:308: in _command
    return conn.command(
python3.11/site-packages/pymongo/helpers.py:322: in inner
    return func(*args, **kwargs)
python3.11/site-packages/pymongo/pool.py:968: in command
    return command(
python3.11/site-packages/pymongo/network.py:192: in command
    helpers._check_command_response(

This seems non-deterministic and re-running the same test will not always encounter this issue. I wonder if there is some delay or error in initialization this service in the local deployment. Is there any advice on how to solve or avoid this?

2

Hi @Produce_AI, welcome to MongoDB community!

Would it be possible to provide us with the following details on the error

  1. Run atlas deployments diagnostics after a failure and upload the logs (see: troubleshooting)
  2. Does this issue happen only when creating a deployment then goes away, and / or happens when the deployment is running for a while?
  3. If possible, share the test code

Thank you!

3

Same issue here. Did you fix it ?

4

Hi,

I encountered the same issue, this time while using .NET, so it doesn’t appear to be driver-related.
I was able to “fix” it by allowing more time for the MongoDB Atlas container to start up — simply adding a delay during test initialization resolved the problem.
Probably not the best solution but it works.