This project is deprecated in favor of: Postgers Importer

Performs a pg_dump against a source database and restores it into the specified target.

1. Clone this project and update the Dockerfile's PG_VERSION to match the version of your target Postgres cluster.

2. Run fly apps create --name postgres-migration-tool

3. Specify the SOURCE_DATABASE_URI and TARGET_DATABASE_URI as secrets.

# The source postgres database.
fly secrets set SOURCE_DATABASE_URI=postgres://<user>:<password>@<hostname>:<port>/<database>

# The target postgres database ( Leave off the database name unless the database already exists ).
fly secrets set TARGET_DATABASE_URI=postgres://<user>:<password>@<hostname>:<port>

4. Deploy app

fly deploy

5. Read the recommendations section below. :)

6. Initiate the migration process

fly ssh console -C "migrate"

• Make sure the target app has plenty of memory.
• Scale your destination app down to 1 before initiating the migration: fly scale count 1 -a target-db-name. It might warn you that this app name is different than what's in the current fly.toml, but that's ok
• If the migration looks like it has frozen, connect to your target postgres app and run \l+ a few times to make sure the data size is still growing.
• If migration connection is broken, your target app is likely under scaled. Monitor your metrics dashboard and your Fly app's vm checks for guidance.