Managing Postgres access

Suggest edits

You control access to your Postgres database using database authentication implemented by creating databases with specific roles and privileges. Database authentication differs from portal authentication, which controls access to the BigAnimal portal.

For information on portal authentication, see:

Setting up your database authentication

Don't use the edb_admin database role and edb_admin database created when creating your cluster in your application. Instead, create a new database role and a new database, which provides a high level of isolation in Postgres. If multiple applications are using the same cluster, each database can also contain multiple schemas, essentially a namespace in the database. If you need strict isolation, use a dedicated cluster or dedicated database. If you don't need that strict isolation level, you can deploy a single database with multiple schemas. See Privileges in the PostgreSQL documentation to further customize ownership and roles to your requirements.

To create a new role and database, first connect using psql:

psql -W "postgres://edb_admin@xxxxxxxxx.xxxxx.biganimal.io:5432/edb_admin?sslmode=require"
Note

Avoid storing data in the postgres system database.

Admin roles

pg_ba_admin

So that we can effectively manage the cloud resources and ensure users are protected against security threats, BigAnimal provides a special administrative role, pg_ba_admin. The edb_admin user is a member of the pg_ba_admin role. The pg_ba_admin role has privileges similar to a Postgres superuser. Like the edb_admin user, the pg_ba_admin role shouldn't be used for day-to-day application operations and access to the role must be controlled carefully. See pg_ba_admin role for details.

superuser

When using your own cloud account, you can grant the edb_admin role superuser privileges for a cluster. See Superuser access. If you grant superuser privileges, you must take care to limit the number of connections used by superusers to avoid degraded service or compromising availability.

Note

Superuser privileges allow you to make Postgres configuration changes using ALTER SYSTEM queries. We recommend that you don't do this because it might lead to an unpredictable or unrecoverable state of the cluster. In addition, ALTER SYSTEM changes aren't replicated across the cluster.

See the PostgreSQL documentation on superusers for best practices.

Notes on the edb_admin role

  • Changes to system configuration (GUCs) made by edb_admin or other Postgres users don't persist through a reboot or maintenance. Use the BigAnimal portal to modify system configuration.

  • You have to remember your edb_admin password, as EDB doesn't have access to it. If you forget it, you can set a new one in the BigAnimal portal on the Edit Cluster page.

  • Don't use the edb_admin user or the edb_admin database in your applications. Instead, use CREATE USER; GRANT; CREATE DATABASE.

  • BigAnimal stores all database-level authentication securely and directly in PostgreSQL. The edb_admin user password is SCRAM-SHA-256 hashed prior to storage. This hash, even if compromised, can't be replayed by an attacker to gain access to the system.

One database with one application

For one database hosting a single application, replace app1 with your preferred user name:

  1. Create a new database user. For example,

    edb_admin=# create user app1 with password 'app1_pwd';

  2. Assign the new role to your edb_admin user. Assigning this role allows you to assign ownership to the new user in the next step. For example:

    edb_admin=# grant edb_admin to app1;

  3. Create a new database to store application data. For example:

    edb_admin=# create database app1 with owner app1;

Using this example, the username and database in your connection string is app1.

One database with multiple schemas

If you use a single database to host multiple schemas, create a database owner and then roles and schemas for each application. This example shows creating two database roles and two schemas. The default search_path for database roles in BigAnimal is "$user",public. If the role name and schema match, then objects in that schema match first, and no search_path changes or fully qualifying of objects are needed. The PostgreSQL documentation covers the schema search path in detail.

  1. Create a database owner and new database. For example:

    edb_admin=# create user prod_admin with password 'prod_pwd';
edb_admin=# create database prod1 with owner prod_admin;

  2. Connect to the new database. For example:

    edb_admin=# \c prod1

  3. Create new application roles. For example:

    prod1=# create user app1 with password 'app1_pwd';
prod1=# grant edb_admin to app1;

prod1=# create user app2 with password 'app2_pwd';
prod1=# grant edb_admin to app2;

  4. Create a new schema for each application with the AUTHORIZATION clause for the application owner. For example:

    prod1=# create schema app1 authorization app1;
prod1=# create schema app2 authorization app2;

IAM authentication for Postgres

Any AWS user with an AWS account connected to a BigAnimal subscription who has the Postgres role of iam_aws can authenticate to the database using their AWS IAM credentials.

Configuring IAM for Postgres

Provision your cluster before configuring IAM for Postgres.

  1. In BigAnimal, turn on the IAM authentication feature when creating or modifying the cluster:

    1. On the Additional Settings tab, under Authentication, select Identity and Access Management (IAM) Authentication.
    2. Select Create Cluster or Save.
      Note

      To turn on IAM authentication using the CLI, see Using IAM authentication on AWS.

  2. In AWS, get the ARN of each IAM user requiring database access. In the AWS account connected to BigAnimal, use AWS Identity and Access Management (IAM) to perform user management. See the AWS documentation.

  3. In Postgres, if the IAM role doesn’t exist yet, run this Postgres command:

    CREATE ROLE "iam_aws";

  4. For each IAM user, run this Postgres command:

    CREATE USER "<ARN>" IN ROLE iam_aws;

Logging in to Postgres using IAM credentials

If IAM integration is configured for your cluster, you can log in to Postgres using your AWS Amazon Resource Name (ARN) and access key. Using this ARN + access key combination allows you to connect to your Postgres database using your AWS IAM standard credentials.

You can continue to log in using your Postgres username and password. However, doing so doesn’t provide IAM authentication even if this feature is configured.

  1. Using your AWS CLI or Cloud shell, obtain your ARN and access key. For guidance on obtaining your ARN and access key, see Managing access keys for IAM users.
  2. Connect to Postgres using your IAM credentials.
  3. When prompted for the password, enter your access key (<access key ID><secret access key>).

Using IAM authentication CLI commands

For information on integrating with IAM on AWS using the CLI, see IAM authentication CLI commands.

← Prev

Using your cluster

Next →

pg_ba_admin role