Active-Passive Setup

An active-passive setup provides redundancy by maintaining a secondary (passive) server that can take over if the primary (active) server fails. This guide explains how to configure a Surfly active-passive setup.

Implementation Steps

1. Install Surfly on Passive Node

Install the exact same version of Surfly on the passive node by following the installation documentation. Ensure all configuration steps are identical except for any host-specific settings.

2. Create Database Dumps on Active Node

On the active node, create dumps of both the Surfly and Console databases:

# Create Surfly database dump
pg_dump -U surfly_app surfly > ~/surfly_$(date +"%Y-%m-%d").sql

# Create Console database dump
pg_dump -U surfly_app console > ~/console_$(date +"%Y-%m-%d").sql

3. Transfer Database Dumps to Passive Node

Copy the database dumps from the active to the passive node:

# From active node, replace PASSIVE_NODE_IP with your passive node's IP address
scp ~/surfly_*.sql ~/console_*.sql user@PASSIVE_NODE_IP:/tmp/

4. Import Databases on Passive Node

On the passive node, stop required services before database restoration:

systemctl --user stop ss-pgbouncer
sudo systemctl restart postgresql

Import the databases:

# For SQL files
dropdb -U postgres surfly && createdb -U postgres -O surfly_app -E UTF8 surfly && psql -U surfly_app surfly < /tmp/surfly_YYYY-MM-DD.sql
dropdb -U postgres console && createdb -U postgres -O surfly_app -E UTF8 console && psql -U surfly_app console < /tmp/console_YYYY-MM-DD.sql

Restart services after restoration:

systemctl --user start ss-surfly.target

Manual Failover

To manually switch to the passive node:

Update your load balancer configuration to direct traffic to the passive node
Verify the passive node is operating correctly

Automatic Failover

For automatic failover, configure a load balancer as described in the Failover documentation.