Mongodb/atlas docker container - unable to start deployment if the container restart

[Jose_Salina]

I created this container following this documentation

mongo:
    container_name: mongo
    image: mongodb/atlas
    privileged: true
    command: |
        /bin/bash -c "atlas deployments setup MyLocalContainer --type local --port 27017 --bindIpAll --mdbVersion 6.0 --username root --password root --force && tail -f /dev/null"
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
    ports:
      - "27017:27017"
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock

The first build was success and I was able to connect using this string

mongodb://root:root@localhost:27017/leadconduit_local?directConnection=true&authSource=admin

but when the container is restarted, I got this error message and I’m unable to connect again

Creating your cluster MyLocalContainer

1/2: Starting your local environment...

Error: "MyLocalContainer" deployment already exists and is currently in "running" state

Error: currently there are no deployments

Connection string: mongodb://root:root@localhost:27017/?directConnection=true

I tried removing the deployment name but every time that the container restarts it create another deployment and I should re-setup again.

Is there a way to keep the deployment name?

[John_Bench]

I have encountered the same issue. If I try to start the container again after that first error (Error: "MyLocalContainer" deployment...) Jose posted I get the following:

Creating your cluster MyLocalContainer
2024-02-06T19:11:38.382999473Z 1/2: Starting your local environment...
2024-02-06T19:11:38.425274647Z 2/2: Creating your deployment MyLocalContainer...
2024-02-06T19:11:38.587681695Z Error: exit status 125: Error: network name mdb-local-MyLocalContainer already used: network already exists

[John_Bench]

Would this be fixed by having an updated volume mount point other than /data/db? If so, what should the new mount point be?

[Brayden_Tidwell]

We’re experiencing similar issues. I’ve found a work around for our case that may help you:

Introducing a Local Experience for Atlas, Atlas Search, and Atlas Vector Search with the Atlas CLI

Persistency is a major setback to our team as well. If anyone managed to solve it, please share

[John_Bench]

Awesome! Thank you @Brayden_Tidwell : I was able to use that entrypoint methodology to address our issue.

Here is what I wound up using to be able to restart the container and preserve the data after restart, stop, and compose down:
start-atlas.sh

#!/bin/bash

cleanup () {
  # stop service and clean up here
  echo "stopping atlas deployment"
  atlas deployments stop MyMongo --type local 2>/dev/null
  echo "stopping podman containers"
  podman stop --all

  exit 0
}

trap "cleanup" EXIT
# 2>/dev/null is used to silence output about listing atlas instances other than local
PODMAN_HAS_MONGO_CONTAINER=$(podman ps --all 2>/dev/null | grep 'MyMongo')
PODMAN_HAS_MONGO_NETWORK=$(podman network ls  2>/dev/null | grep 'mdb-local-MyMongo')
DEPLOYMENT_INFO=$(atlas deployments list  2>/dev/null | grep 'MyMongo')

if [[ $PODMAN_HAS_MONGO_CONTAINER ]]; then
    # If missing network, create it (happens after docker compose down)
    if ! [[ $PODMAN_HAS_MONGO_NETWORK ]]; then
      # silence the update check
      atlas config set skip_update_check true 2>/dev/null
      echo "creating podman network:"
      podman network create mdb-local-MyMongo
    fi
    # Restart a deployment
    echo "starting podman containers"
    podman start --all
fi

if [[ $DEPLOYMENT_INFO ]]; then
    atlas deployments start MyMongo --type local 2>/dev/null
else
    # silence the update check
    atlas config set skip_update_check true 2>/dev/null
    atlas deployments setup MyMongo --type local --username root --password root --port 27017 --bindIpAll --force 2>/dev/null
fi

sleep infinity &
wait $!

docker-compose.yml

  my_mongo:
    container_name: my_mongo
    image: mongodb/atlas:v1.14.2
    privileged: true
    ports:
      - "27017:27017"
    entrypoint: /home/scripts/start-atlas.sh
    volumes:
      - ./scripts/start-atlas.sh:/home/scripts/start-atlas.sh
      - /var/run/docker.sock:/var/run/docker.sock # <---- didn't really find this to be necessary included because it was in the orignal docs
      - mongodb-data:/var/lib/containers # <---- needed to perserve podman containers for restart after docker compose down

I was also able to use the container in our GitHub Actions by using the following action:

      - name: Docker Compose
        uses: isbang/compose-action@v1.5.1
        with:
          services: |
            my_mongo
          up-flags: "--build --detach --pull always --quiet-pull --wait --wait-timeout 300"
          down-flags: "--volumes --remove-orphans --timeout 5"