Database Cluster (baboon.sshjunkie.com)

Overview

The database cluster consists of two PostgreSQL database servers hosted on baboon.sshjunkie.com. These servers are used to store data for services such as Mastodon and AzuraCast. The cluster ensures high availability and fault tolerance through replication and backup strategies.

Installation

Install PostgreSQL on both nodes in the cluster:

# Update package list and install PostgreSQL
sudo apt update
sudo apt install -y postgresql postgresql-contrib

# Ensure PostgreSQL is running
sudo systemctl start postgresql
sudo systemctl enable postgresql

Configuration

PostgreSQL Configuration Files:

pg_hba.conf:

Allow replication and local connections.

Example:

local   all             postgres                                md5
host    replication     all             192.168.0.0/16            md5

postgresql.conf:

Set wal_level for replication:

wal_level = hot_standby
max_wal_senders = 3

Replication Configuration:

Set up streaming replication between the two nodes (baboon.sshjunkie.com as the master and the second node as the replica).

On the master node, enable replication and restart PostgreSQL.
On the replica node, set up replication by copying the data directory from the master node and configure the recovery.conf file.

Example recovery.conf on the replica:

standby_mode = on
primary_conninfo = 'host=baboon.sshjunkie.com port=5432 user=replicator password=your_password'
trigger_file = '/tmp/postgresql.trigger.5432'

Usage

Check the status of PostgreSQL:
```
sudo systemctl status postgresql
```

Promote the replica to master:

pg_ctl promote -D /var/lib/postgresql/data

Backups

Use pg_basebackup to create full backups of the cluster. Example:

pg_basebackup -h baboon.sshjunkie.com -U replicator -D /backups/db_backup -Ft -z -P

Automate backups with cronjobs for regular snapshots.

Troubleshooting

Issue: Replica is lagging behind.
- Solution: Check network connectivity and ensure the replica is able to connect to the master node. Monitor replication lag with:
```
SELECT * FROM pg_stat_replication;
```

Monitoring

Monitor replication status:
```
SELECT * FROM pg_stat_replication;
```
Monitor database health:
```
pg_isready
```

Additional Information

PostgreSQL Streaming Replication Documentation

2.5 KiB Raw Blame History