This application runs alongside Mongo to create a simple API endpoint that can be called for backup operations.

curl -XPOST 'http://localhost:8080/v0/dump' -d \
'{
    "bucket": "org-mongodb-snapshots",
    "path": "/dev/"
}'

This creates a timestamped file in your S3 bucket at the path you specify. This file is compatible with mongorestore, using the --archive option

I wanted to be able to use a Lambda scheduled job to back up Mongo to S3, but using mongodump in python required more than 1500MB of memory on my dataset and was very slow.

If mongo is running in another container:

docker pull openwhere/go-mongo-backup
docker run --name mongo-backup --link mongo -d -p 8080:8080 openwhere/go-mongo-backup

If mongo is running as a native package on the host:

docker pull openwhere/go-mongo-backup
docker run --name mongo-backup -d -p 8080:8080 \
  -e "MONGOHOST=$(ifconfig | grep -E '([0-9]{1,3}\.){3}[0-9]{1,3}' | grep -v 127.0.0.1 | awk '{ print $2 }' | cut -f2 -d: | head -n1)" \
  openwhere/go-mongo-backup

This sets the mongo address to the IP of the Docker host.

To start a backup:
POST /v0/dump

{
    "bucket": "mybucket",
    "path": "/path/to/backup/dir/"
}

1. Test the lambda function python lambda/mongo-backup/mongo-snapshot.py
2. create zip ./lambdabuild.sh
3. Create Lambda function using zip file. (Automated in pod config)

ToDo

Pull requests accepted.

MIT