Proper deployment configuration is critical to a secure solution. This chapter contains detailed deployment recommendations.
While it would have been possible to make Jana work with other databases, there is only a single one supported: PostgreSQL. In order to deploy Jana, the following steps need to be followed. Unlike most instructions, these will guide you through using the psql command-line interface (CLI), rather than the administrative commands.
First, you need to create two distinct users in PostgreSQL. The easiest way to do this is to launch psql as a database superuser:
$ psql -U postgres psql (9.4.0) Type "help" for help. postgres=#
From here, we can create the users we need:
CREATE USER jana_admin WITH LOGIN ENCRYPTED PASSWORD '<something>'; CREATE USER jana WITHLOGIN ENCRYPTED PASSWORD '<somethingdifferent>';
In each case, you should replace the <something> or <somethingdifferent> with the actual passwords. These should be as strong as possible and not guessable.
Next, we need to create the database:
CREATE DATABASE jana ENCODING 'UTF8';
This will create a UTF-8 encoded database. This allows us to safely store strings with non-ASCII characters in them.
The schema for Jana is managed through controlled migrations. These reflect incremental changes to the database that mirror the underlying code development. They are managed using the Alembic library.
One of the things you should do once the initial tables are created is adjust the permissions. Unfortunately, the standard migrations and table creation scripts assume a single user with full access to the database. This isn’t the best way to handle things from a security perspective. Instead, we want to remove some permissions from certain tables from the normal “session” user, jana. Once again, at a superuser prompt for psql, we want to revoke permissions:
REVOKE ALL PRIVILEGES ON TABLE audit FROM jana;
Now, we want to restore just the permissions we want the user to have:
GRANT SELECT ON TABLE audit TO jana; GRANT INSERT ON TABLE audit TO jana;