Skip to content

Latest commit

 

History

History
83 lines (51 loc) · 3.79 KB

README.md

File metadata and controls

83 lines (51 loc) · 3.79 KB
Raw

trek

Trek is a tool that assists in PostgreSQL schema development and deployment. Trek provides a unified and simple workflow for developers to change the database schema and commit versioned migration files to a VCS such as git. Trek also helps with rolling out those migrations to a live database.

A basic trek workflow looks like this:

  • Initialize a trek directory mkdir my_database; cd my_database; trek init
  • Change the database schema by using the pgModeler GUI.
  • Run trek to generate a new migration file. The migration file will contain all changes since the last version as PostgreSQL DDL (ALTER TABLE, etc.). The migration file is stored as file in the my_database/migrations folder.
  • Commit the migration file to an VCS like git. The changed pgModeler schema and the migration SQL file may be code-reviewed.
  • Deploy the latest version of my_database to a live database by running trek apply. Trek uses golang-migrate to keep track of the latest version deployed to a database, and to deploy new migration files to the database. (You may skip trek apply and integrate with golang-migrate manually.)

A developer could run trek generate --dev --stdout to continuously watch the pgModeler model for updates (on save) and write PostgreSQL DDL to stdout.

Trek is an opinionated tool. We built it to do one thing and do it well: convert pgModeler schema design into versioned migrations. Trek is exclusively designed for use with PostgreSQL, and we firmly intend to maintain this focus. As such, we will not be considering or adding support for any other database systems.

Compatibility

Trek works with PostgreSQL version 13 or newer. We build trek with the latest stable version of pgModeler.

We develop and test Trek on Linux. Trek may work on other platforms via Docker.

Installation

go install github.com/printeers/trek@latest

Or use a released docker image.

Dependencies

Running trek generate requires to have migra and pgmodeler-cli installed. Trek will try to locate these in the $PATH of the user.

If migra cannot be found, trek will try use versions of this binary that is embedded into the trek build. The user can skip searching for migra in the path by setting TREK_FORCE_EMBEDDED_MIGRA. For example:

trek --force-embedded-migra=true generate --stdout

Note that the embedded migra is built using a patched schemainspect library, which is awaiting upstream merge.

MacOS / Windows

If you're running on MacOS or Windows, then the embedded migra doesn't work. Either install migra manually or use docker to run trek under linux/amd64 (Rosetta emulation).

Docker

To run trek in docker:

cd your_trek_working_directory
docker run -v ./:/data ghcr.io/printeers/trek:latest-pgmodeler trek ...

Setup a trek working directory manually

We recommended to use trek init, but you may setup a working directory manually.

Inside a directory create trek.yaml:

model_name: <model_name>
db_name: <db_name>
db_users:
  - <db_user_1>
  - <db_user_2>

Create <model_name>.dbm using pgModeler.

Generating a new migration

trek generate some-migration

Use the --dev flag to continuously watch for file changes. Use the --stdout flag to write migrations to stdout. You must omit the migration name when using --stdout.

Applying the migrations

Take a look at the example/ directory.

History

trek was originally developed by Stack11. In april 2023 Printeers adopted the project for further development and maintenance.