11 / 11
Feb 14

Using service as described here: https://www.mongodb.com/docs/atlas/cli/current/atlas-cli-deploy-docker/#run-the-image-with-github-actions
in github actions:

services: mongodb: image: mongodb/mongodb-atlas-local ports: - 27017:27017

but it always fails:

starting mongodb service is starting, waiting 2 seconds before checking again. /usr/bin/docker inspect --format="{{if .Config.Healthcheck}}{{print .State.Health.Status}}{{end}}" c52ef7f4023438f1ab52d3ddc69394c62d69922474fde4e22e6f332058b263be starting mongodb service is starting, waiting 3 seconds before checking again. /usr/bin/docker inspect --format="{{if .Config.Healthcheck}}{{print .State.Health.Status}}{{end}}" c52ef7f4023438f1ab52d3ddc69394c62d69922474fde4e22e6f332058b263be unhealthy Service container mongodb failed. Error: Failed to initialize container mongodb/mongodb-atlas-local Error: One or more containers failed to start.

Update: the issue has been fixed, and specifying mongodb/mongodb-atlas-local should work again as expected.


In the meantime, you can pin to an older (working) version of the image by updating the yaml. For example:

image: mongodb/mongodb-atlas-local@sha256:364c10e8de7fade95be8939fc817d15776f3724459ae689d078725c54a941333

If you would like a specific older version to pin to, you can navigate to https://hub.docker.com/r/mongodb/mongodb-atlas-local/tags?name=20250205, select the tag you want, and copy the sha into your yaml in the format example given above.

Indeed, many docker images where pushed (and updated while they should be inmutable) and it seems the issue is wide spread
Currently, we figured out that 7.0.3-20250205T151413Z is still working as it was before these issue happened
Hope this helps other to mitigate

We started to suffer the same issue a few minutes ago, and it happens on Linux based OS, it still works as expected on MacOS based machines

Maybe it’s related, but trying to follow Getting Started with MongoDB Atlas Local Search Experience Using Docker, the mongodb-atlas-local:latest image suggested in the tutorial shutdowns immediately.
Browsing the Mongo’s dockerhub I saw a lot of image updated today, so I tried with a slightly older one (8.0.3-20250205T151413Z) and now the container stays up as expected.
From a cursory glance the only difference from this and the lastest image, apart from the mongo version, is layer 35: ENV JAVA_TOOL_OPTIONS=-XX:UseSVE=0

Hi @Nestor_Lobo ,

while they should be immutable

Just to clarify, is it your expectation that the container image tag be immutable?

We started to suffer the same issue […] it still works as expected on MacOS based machines

Sorry to hear this. Could you provide any more details from logs or --debug output, please?

@John_Smith12 just re-reading the tutorial, and assuming you’re using the docker-compose.yml example, it looks like the example (in the code repo too) is missing the line hostname: mongodb. Without this, when the container is stopped and then restarted, the container will struggle to find the volume again. If this is what happened, you should be able to get back into a working state by

  1. Add the hostname the yml
... mongodb: **hostname: mongodb** image: mongodb/mongodb-atlas-local ...
  1. Delete the container in docker desktop
    Or via the terminal
//get the name of the container docker ps -a //delete the container docker rm <name of container or ID>
  1. Docker Compose up

Importantly, when it comes to stopping the container, use docker compose down -v to ensure the data volume persists (assuming you want it to persist, and don’t want to start from scratch each time).

The container with the fixed image seems to be working fine even without the hostname configuration in the docker-compose; I’ll add it nonetheless.
Thank you.

Closed on Feb 14

This topic was automatically closed 5 days after the last reply. New replies are no longer allowed.