如何在docker-compose启动MongoDB副本集后执行Mongo Shell命令？

Great question! Automating the replica set initialization saves you from manual steps every time you spin up the cluster. Here are a couple of reliable approaches to achieve this:

Approach 1: Use a Temporary Initialization Container (Recommended) #

This approach uses a separate one-off container that waits for all three MongoDB nodes to be fully ready, then runs the replica set initiation command. It keeps your core MongoDB service definitions clean and avoids modifying the base mongo image behavior.

Update your docker-compose.yml like this:

version: "3"
services:
  mongo1:
    image: mongo
    ports:
      - 27017:27017
    command: mongod --replSet rs0
  mongo2:
    image: mongo
    ports:
      - 27018:27017
    command: mongod --replSet rs0
    depends_on:
      - mongo1
  mongo3:
    image: mongo
    ports:
      - 27019:27017
    command: mongod --replSet rs0
    depends_on:
      - mongo1
  init-replica:
    image: mongo
    depends_on:
      - mongo1
      - mongo2
      - mongo3
    command: >
      bash -c "
        # Wait for each MongoDB node to become reachable
        until mongo --host mongo1 --eval 'db.adminCommand(\"ping\")'; do 
          echo 'Waiting for mongo1 to be ready...'
          sleep 2
        done
        until mongo --host mongo2 --eval 'db.adminCommand(\"ping\")'; do 
          echo 'Waiting for mongo2 to be ready...'
          sleep 2
        done
        until mongo --host mongo3 --eval 'db.adminCommand(\"ping\")'; do 
          echo 'Waiting for mongo3 to be ready...'
          sleep 2
        done
        # Initialize the replica set
        mongo --host mongo1 --eval 'rs.initiate({\"_id\":\"rs0\",\"members\":[{\"_id\":0,\"host\":\"mongo1:27017\"},{\"_id\":1,\"host\":\"mongo2:27017\"},{\"_id\":2,\"host\":\"mongo3:27017\"}]});'
        echo 'Replica set initialized successfully!'
      "

How it works: #

• The init-replica container uses the official mongo image, so it has all the tools needed to connect to the cluster.
• depends_on ensures the container starts after the three MongoDB nodes, but we still use until loops to wait for each node to respond to a ping (since depends_on only guarantees container startup, not service readiness).
• Once all nodes are ready, it runs the rs.initiate command against mongo1.
• After execution, the container exits (since it's a temporary task) and won't interfere with your running cluster.

Approach 2: Custom Startup Script for mongo1 #

If you prefer to handle initialization directly within the primary node's container, you can create a bash script that starts MongoDB, waits for all nodes to be ready, then runs the initiation command.

1. Create a file named init-replica.sh in the same directory as your docker-compose.yml:

#!/bin/bash
set -e

# Start MongoDB in the background
mongod --replSet rs0 &
MONGOD_PID=$!

# Wait for mongo1 itself to be ready
until mongo --eval 'db.adminCommand("ping")'; do
  echo "Waiting for mongo1 to start..."
  sleep 2
done

# Wait for mongo2 to be ready
until mongo --host mongo2 --eval 'db.adminCommand("ping")'; do
  echo "Waiting for mongo2 to be ready..."
  sleep 2
done

# Wait for mongo3 to be ready
until mongo --host mongo3 --eval 'db.adminCommand("ping")'; do
  echo "Waiting for mongo3 to be ready..."
  sleep 2
done

# Initialize the replica set
echo "Initializing replica set..."
mongo --eval 'rs.initiate({"_id":"rs0","members":[{"_id":0,"host":"mongo1:27017"},{"_id":1,"host":"mongo2:27017"},{"_id":2,"host":"mongo3:27017"}]})'

# Wait for the MongoDB process to keep the container running
wait $MONGOD_PID

2. Update your docker-compose.yml to mount and run this script for mongo1:

version: "3"
services:
  mongo1:
    image: mongo
    ports:
      - 27017:27017
    volumes:
      - ./init-replica.sh:/init-replica.sh
    command: bash /init-replica.sh
  mongo2:
    image: mongo
    ports:
      - 27018:27017
    command: mongod --replSet rs0
  mongo3:
    image: mongo
    ports:
      - 27019:27017
    command: mongod --replSet rs0

Key Notes: #

• The script starts MongoDB in the background, waits for all nodes to be reachable, then runs the initiation command. The final wait ensures the container stays running (since MongoDB is in the background).
• Make sure the init-replica.sh file has executable permissions: run chmod +x init-replica.sh before starting the cluster.

Important Considerations #

• Idempotency: Running rs.initiate multiple times is safe—MongoDB will just return an error saying the replica set is already initialized, which won't break your cluster.
• Networking: All containers must be on the same Docker network (Docker Compose creates a default network for you, so this is handled automatically).
• Readiness Checks: The ping checks (db.adminCommand("ping")) are a reliable way to verify that each MongoDB node is fully operational and ready to accept connections.

内容的提问来源于stack exchange，提问作者atkayla