Create and autorun a PostgreSQL development database for your Django project
Django-pgrunner creates a local PostgreSQL database for your project and automatically starts the database server when needed. It allows you to develop against a PostgreSQL server, while matching the convenience of a SQLite3 database file.
The local database is a self-contained database created using the PostgreSQL
initdb tool. A separate PostgreSQL server will be started on a custom port.
Additionally, it support creating and activating database snapshots. This allows you to experiment with your data and be confident that you can restore your old data within seconds.
While sqlite3 is great for development, being able to develop against a full-featured database like PostgreSQL has a few huge benefits:
- Your development database will match the database you run in production, eliminating a whole category of bugs, like:
- South migrations failing in production, while working in dev
- Errors in code due to database specific behaviour, like stricter checks on foreign key references.
- Proper checks on
unique_together(sqlite3 does not support these constraints)
- You can use all the great features offered by PostgreSQL, instead of restricting yourself to a low common denominator. Examples:
- Database-side constraint and cascading rules
- Extra data types, like arrays
- Full-text search
- Stored procedures
- PostGIS geo extensions
- JSON features
And with the ease of use this packae offers, there is no good reason not to use it!
Make sure you have PostgreSQL installed. You can either compile from source, or install a binary for your operating system. Check the download page for details. For MacOS X users, I recommend downloading PostgresApp.
The snapshot functionality requires rsync to be installed.
This package depends on the psycopg2 package.
psycopg2 provides the Python database adapter needed to communicate
with PostgreSQL. This adapter is partly written in C, and requires a working
C compiler, PostgreSQL development headers and the
pg_config binary in your
PATH. Alternatively, you can install a binary distribution.
pip install psycopg2
The easiest way to install
django-pgrunner is using pip:
pip install django-pgrunner
Microsoft Windows is currently not supported.
This package is not intended for use on production servers.
First, add pgrunner to your INSTALLED_APPS.
Next, add the following lines to your
settings.py or local settings file.
Make sure that these come after any DATABASES setting:
# If your PostgreSQL binaries are not in your path, add this setting #PGRUNNER_BIN = '/usr/lib/postgresql/9.3/bin' #PGRUNNER_BIN = '/Applications/Postgres.app/Contents/Versions/9.3/bin' # This will overwrite DATABASES and auto-start PostgreSQL if needed import pgrunner pgrunner.settings(locals())
Note that most Linux distibutions do not include these binaries in the PATH,
and only expose a few wrappers in
/usr/bin. If this is the case for you, you
need to set the
Run the following command to create your local database:
To start the database in the background:
./manage.py pg_ctl start
Note that if you use pgrunner.setting(), it will automatically start the server for you.
To stop the database:
./manage.py pg_ctl stop
This will not be done automatically.
To start the
psql command line interface with the right parameters, use one of
./manage.py dbshell ./manage.py pg_psql
The only difference is that the first one uses your
DATABASES settings, and the
second one ignores it.
Snapshots are nothing more than named copies of the full database.
It’s the equivalent of a
cp dev.sqlite my-backup.sqlite for SQLite users.
To create a snapshot and activate it:
./manage.py pg_snapshot my-snapshot ./manage.py pg_activate my-snapshot
The name of the default snapshot you are running is
default, so to switch
./manage.py pg_activate default
Snapshots can be deleted by removing their folder under
./manage.py pg_init - Initialize a local database environment ./manage.py pg_ctl start - Start server in background ./manage.py pg_ctl stop - Stop server in background ./manage.py pg_ctl status - Check if the server is running ./manage.py pg_run - Run PostgreSQL server in foreground ./manage.py pg_psql - Start psql with right parameters ./manage.py pg_snapshot foo - Create a copy of all current database data ./manage.py pg_activate - List all snapshots ./manage.py pg_activate foo - Activate snapshot 'foo' ./manage.py pg_dump - Run pg_dump ./manage.py pg_restore - Run pg_restore
Django-pgrunner creates a
pgrunnerdb/ subfolder under your Django project.
This folder contains one folder for every snapshot that was created.
The name of the default instance is
current symbolic link
keeps track of which snapshot is active.
A separate PostgreSQL daemon is started for every project. When the
management command is run, it will pick a random port between 15000 and 16000
to run the server on and write it to the local
postgres.conf. This way
the chances of a conflict between projects are small.