docker-mysqldump

A simple Docker container to dump the contents of a MySQL database (using mysqldump) and copy it into S3.

Using It

Build the image

docker build -t dump .

Run the image

docker run -it --rm --name dump -e "DB_USER=root" -e "DB_PASSWORD=XXXXXXX" -e "DB_HOST=mydb.XXXXX.us-east-1.rds.amazonaws.com" -e "AWS_PROFILE=my-aws-profile" -e "BUCKET=my-project-bucket" -e "DB_NAME=my-db" -v ~/.aws:/root/.aws dump

• Environment Variables Required
  • DB_USER - MySQL user that has enough privileges to mysqldump the target MySQL database.
  • DB_PASSWORD - the password for the DB_USER
  • DB_HOST - FQDN or IP address of the MySQL instance
  • DB_NAME - a label to associate with this backup
  • BUCKET - target S3 bucket
• AWS CLI Environment Variables
  • Any of the methods of specifying AWS CLI credentials (as discussed in AWS documentation). E.g., AWS_PROFILE or AWS_ACCESS_KEY_ID, etc.

Implementation Notes

• Dumps use the "--all-databases" option of mysqldump.
• Dumps are gzipped.
• The S3 bucket is created if it does not exist. "us-east-1" region is assumed.
• Dumps are stored in s3://BUCKET/mysql-dump/DB_NAME/mysql-dump.YYYY-MM-DD-HH-MM.sql.gz.
• Whatever AWS credentials are supplied, they must have permission to put objects into the bucket.

Future Improvements

• Allow specifying the region to use for the bucket, if it is created.
• Provide a "dry-run" option to testing MySQL permissions and S3 permissions without actually doing the export.
• Provide a mechanism to prune older dumps.
• Provide option to use a storage class other than S3 standard.