Postgres Operations

The PostgreSQL modules manage PostgreSQL databases, users and privileges.

Requires the psql CLI executable on the target host(s).

All operations in this module take four optional arguments:

  • psql_user: the username to connect to postgresql to

  • psql_password: the password for the connecting user

  • psql_host: the hostname of the server to connect to

  • psql_port: the port of the server to connect to

  • psql_database: the database on the server to connect to

See example/postgresql.py for detailed example

Facts used in these operations: postgres.PostgresDatabases, postgres.PostgresRoles.

postgres.database

Add/remove PostgreSQL databases.

postgres.database(
    database: str, present=True, owner: str | None=None, template: str | None=None,
    encoding: str | None=None, lc_collate: str | None=None, lc_ctype: str | None=None,
    tablespace: str | None=None, connection_limit: int | None=None,
    psql_user: str | None=None, psql_password: str | None=None, psql_host: str | None=None,
    psql_port: int | None=None, psql_database: str | None=None, **kwargs,
)

  • name: name of the database

  • present: whether the database should exist or not

  • owner: the PostgreSQL role that owns the database

  • template: name of the PostgreSQL template to use

  • encoding: encoding of the database

  • lc_collate: lc_collate of the database

  • lc_ctype: lc_ctype of the database

  • tablespace: the tablespace to use for the template

  • connection_limit: the connection limit to apply to the database

  • psql_*: global module arguments, see above

Updates:

pyinfra will change existing databases - but some parameters are not changeable (template, encoding, lc_collate and lc_ctype).

Example:

postgresql.database(
    name="Create the pyinfra_stuff database",
    database="pyinfra_stuff",
    owner="pyinfra",
    encoding="UTF8",
    _sudo_user="postgres",
)
Note:

This operation also inherits all global arguments.

postgres.dump

Stateless operation

This operation will always execute commands and is not idempotent.

Dump a PostgreSQL database into a .sql file. Requires pg_dump.

postgres.dump(
    dest: str, psql_user: str | None=None, psql_password: str | None=None,
    psql_host: str | None=None, psql_port: int | None=None, psql_database: str | None=None,
    **kwargs,
)

  • dest: name of the file to dump the SQL to

  • psql_*: global module arguments, see above

Example:

postgresql.dump(
    name="Dump the pyinfra_stuff database",
    dest="/tmp/pyinfra_stuff.dump",
    sudo_user="postgres",
)
Note:

This operation also inherits all global arguments.

postgres.load

Stateless operation

This operation will always execute commands and is not idempotent.

Load .sql file into a database.

postgres.load(
    src: str, psql_user: str | None=None, psql_password: str | None=None,
    psql_host: str | None=None, psql_port: int | None=None, psql_database: str | None=None,
    **kwargs,
)

  • src: the filename to read from

  • psql_*: global module arguments, see above

Example:

postgresql.load(
    name="Import the pyinfra_stuff dump into pyinfra_stuff_copy",
    src="/tmp/pyinfra_stuff.dump",
    sudo_user="postgres",
)
Note:

This operation also inherits all global arguments.

postgres.role

Add/remove PostgreSQL roles.

postgres.role(
    role: str, present: bool=True, password: str | None=None, login: bool=True,
    superuser: bool=False, inherit: bool=False, createdb: bool=False, createrole: bool=False,
    replication: bool=False, connection_limit: int | None=None, psql_user: str | None=None,
    psql_password: str | None=None, psql_host: str | None=None, psql_port: int | None=None,
    psql_database: str | None=None, **kwargs,
)

  • role: name of the role

  • present: whether the role should be present or absent

  • password: the password for the role

  • login: whether the role can login

  • superuser: whether role will be a superuser

  • inherit: whether the role inherits from other roles

  • createdb: whether the role is allowed to create databases

  • createrole: whether the role is allowed to create new roles

  • replication: whether this role is allowed to replicate

  • connection_limit: the connection limit for the role

  • psql_*: global module arguments, see above

Updates:

pyinfra will not attempt to change existing roles - it will either create or drop roles, but not alter them (if the role exists this operation will make no changes).

Example:

postgresql.role(
    name="Create the pyinfra PostgreSQL role",
    role="pyinfra",
    password="somepassword",
    superuser=True,
    login=True,
    _sudo_user="postgres",
)
Note:

This operation also inherits all global arguments.

postgres.sql

Stateless operation

This operation will always execute commands and is not idempotent.

Execute arbitrary SQL against PostgreSQL.

postgres.sql(
    sql: str, psql_user: str | None=None, psql_password: str | None=None,
    psql_host: str | None=None, psql_port: int | None=None, psql_database: str | None=None,
    **kwargs,
)

  • sql: SQL command(s) to execute

  • psql_*: global module arguments, see above

Note:

This operation also inherits all global arguments.