Backup and Recovery: Single-Server Streaming - Configuring PostgreSQL

In this step, we'll configure a PostgreSQL server, preparing it to be backed up using Barman.

For Barman to back up this server, a few things need to be done to prepare it:

1. Installing the Barman CLI tools
2. Creating a dedicated superuser for Barman to connect as
3. Creating a dedicated streaming user with the replication attribute and remote login permissions
4. Ensuring there are free replication slots

Barman CLI installation

We'll start by installing the barman-cli package: this contains the barman-wal-archive and barman-wal-restore commands that will be used to transmit data to and from our backup server.

Since PostgreSQL is already installed, the PostgreSQL apt repository is already configured and we can just request the package:

sudo apt-get update
sudo apt-get -y install barman-cli

...
Setting up barman-cli-cloud (2.12-1.pgdg100+1) ...
Processing triggers for ca-certificates (20200601~deb10u2) ...
Updating certificates in /etc/ssl/certs...
0 added, 0 removed; done.
Running hooks in /etc/ca-certificates/update.d...
done.

User provisioning

Let's call our dedicated backup user, "barman". We'll create it interactively via the createuser utility:

createuser --superuser --replication -P barman

Enter password for new role:
Enter it again:

Enter example-password when prompted (twice).

We're making this a superuser account, which will allow it full control of all databases on this server. Be very careful with superuser credentials! Anyone who can connect as a superuser account owns your data.

Now we will create the streaming replication user, "streaming_barman". This doesn't need to be a superuser, but it does need the replication attribute:

createuser --replication -P streaming_barman

Enter password for new role:
Enter it again:

Enter example-password when prompted (twice).

We'll also need to edit pg_hba.conf to allow the streaming user to connect from the backup server, by explicitly allowing it to connect in replication mode:

sed -i '$ a host   replication    streaming_barman   all md5' /var/lib/postgresql/data/pg_hba.conf
cat /var/lib/postgresql/data/pg_hba.conf

...
local   replication     all                                     trust
host    replication     all             127.0.0.1/32            trust
host    replication     all             ::1/128                 trust

host all all all md5
host   replication    streaming_barman   all md5

Enable remote access to your PostgreSQL Server

If the PostgreSQL server isn't already configured for remote access, or is restricted to connections from known machines, our database server won't be able to connect.

We'll use psql to check this setting:

psql -d pagila

The relevant configuration parameter is listen_addresses:

show listen_addresses;

 listen_addresses 
------------------
 *
(1 row)

The value for this server is * - which allows connections from anything. This doesn't need to be changed. If the value were to be empty, localhost, or a list of hosts or addresses that don't include our database server, we'd need to add its hostname (backup) to the list, or change it to the wildcard.

ALTER SYSTEM SET listen_addresses TO "*";

ALTER SYSTEM

Database settings for streaming

We'll also need to make sure there are replication slots available, and that PostgreSQL will allow another sender to connect. We'll continue to use psql to check the current settings:

Show max_wal_senders;
Show max_replication_slots;

 max_wal_senders
-----------------
 10
(1 row)

 max_replication_slots
-----------------------
 10
(1 row)

The default for both of these (for PostgreSQL 10 and above) is 10, so we're fine - but if we needed more (or if they'd been previously set to 0, thus disabling replication) then we'd need to increase them. You could do this by using the commands ALTER SYSTEM SET max_wal_senders = 10; and ALTER SYSTEM SET max_replication_slots = 10;.

Gazing fondly at data

As we mentioned at the start, the server is preloaded with the Pagila database. This is a schema designed for examples and includes, among other things, information about actors. Before we end this step, let's query some of that data - this is what we're going to back up!

select * from actor where last_name='KILMER';
\q

 actor_id | first_name | last_name |      last_update
----------+------------+-----------+------------------------
       23 | SANDRA     | KILMER    | 2020-02-15 09:34:33+00
       45 | REESE      | KILMER    | 2020-02-15 09:34:33+00
       55 | FAY        | KILMER    | 2020-02-15 09:34:33+00
      153 | MINNIE     | KILMER    | 2020-02-15 09:34:33+00
      162 | OPRAH      | KILMER    | 2020-02-15 09:34:33+00
(5 rows)

We'll verify later on that this can be restored reliably.

Continue on with Step #2: Backup Server Configuration.