Active-Passive Topology

Active-Passive topology is a special mode where replication-manager monitors a single active database server without managing replication or performing automatic failover. This topology is useful for scenarios requiring external orchestration or when automatic failover is not desired.

Configuration

replication-active-passive (3.1)

Topology Characteristics

Behavior

In Active-Passive mode:

• Single active server: Only one server is designated as "master"
• No replication management: replication-manager does not configure or manage replication
• No automatic failover: Failover and switchover operations are disabled
• No rejoin operations: Server rejoin is skipped
• No privilege checks: REPLICATION and SUPER privilege checks are bypassed
• Simplified monitoring: Focuses on server availability and health metrics

Automatic Detection

Active-Passive topology is automatically enabled when:

• Configuration explicitly sets replication-active-passive = true
• Only one server is defined in the db-servers-hosts list

Use Cases

1. External Orchestration

Use Active-Passive mode when failover is managed by external systems:

• Kubernetes operators
• Cloud provider managed databases (RDS, Cloud SQL, Azure Database)
• External clustering solutions (Pacemaker, Corosync)
• Manual failover procedures

2. Development/Testing

Monitor a standalone database without replication complexity:

• Local development environments
• Integration testing
• Database migration testing

3. Single-Server Production

Monitor production databases that don't require high availability:

• Low-traffic applications
• Non-critical workloads
• Budget-constrained deployments

4. Monitoring-Only Deployments

Collect metrics and health information without infrastructure changes:

• Performance monitoring
• Backup scheduling
• Schema change detection
• Query analysis

Features Available in Active-Passive Mode

Active-Passive mode supports:

• Monitoring: Server health, status variables, processlist
• Metrics collection: Graphite, Prometheus endpoints
• Backup management: Logical and physical backups via restic
• Schema monitoring: Track schema changes and differences
• Query monitoring: Long queries, Performance Schema analysis
• API access: Full REST API for server information
• Web UI: Dashboard with server metrics and status

Features Disabled in Active-Passive Mode

Active-Passive mode disables:

• Failover: Manual and automatic failover operations blocked
• Switchover: Cannot perform switchover operations
• Rejoin: Automatic server rejoin disabled
• Replication setup: No CHANGE MASTER commands issued
• Read-only enforcement: Does not set read_only flag
• Privilege checks: Skips replication user privilege validation
• Maintenance mode: Maintenance flag not applied to master

Configuration Example

Minimal Configuration

[default]
db-servers-hosts = "192.168.1.100:3306"
db-servers-credential = "root:password"
replication-active-passive = true
monitoring-datadir = "/var/lib/replication-manager"

Full Monitoring Configuration

[myapp]
db-servers-hosts = "db.example.com:3306"
db-servers-credential = "monitor:password"
replication-credential = "replication:password"
replication-active-passive = true

# Monitoring
monitoring-datadir = "/var/lib/replication-manager/myapp"
monitoring-write-heartbeat = false
graphite-metrics = true
graphite-embedded = true

# Backups
backup-restic = true
backup-restic-repository = "s3:https://s3.amazonaws.com/my-backups"
backup-keep-daily = 7
backup-keep-weekly = 4

# Schema monitoring
monitoring-schema-change = true
monitoring-schema-change-script = "/usr/local/bin/schema-alert.sh"

# Query monitoring
monitoring-queries = true
monitoring-long-query-time = 5000

Multi-Server Active-Passive (No Replication)

[cluster1]
db-servers-hosts = "server1:3306,server2:3306,server3:3306"
db-servers-credential = "root:password"
replication-active-passive = true

When multiple servers are configured with Active-Passive mode:

• First reachable server becomes "master"
• Other servers are marked as "Unconnected"
• No replication is expected or configured
• Useful for monitoring multiple independent databases as a single logical cluster

Transitioning Between Modes

From Master-Slave to Active-Passive

To disable automatic failover on an existing replicated cluster:

replication-active-passive = true

replication-manager will:

• Stop performing failover operations
• Continue monitoring existing replication status
• Maintain current topology unchanged
• Allow manual intervention for topology changes

From Active-Passive to Master-Slave

To enable automatic failover:

1. Ensure replication is configured on database servers
2. Update configuration:

replication-active-passive = false

3. Restart replication-manager or reload configuration
4. replication-manager will detect topology and enable failover

API Behavior

In Active-Passive mode, API endpoints behave differently:

Blocked endpoints:

• POST /api/clusters/{cluster}/actions/switchover - Returns error
• POST /api/clusters/{cluster}/actions/failover - Returns error
• POST /api/clusters/{cluster}/servers/{server}/actions/rejoin - Returns error

Available endpoints:

• All GET endpoints for monitoring
• Backup management endpoints
• Server maintenance endpoints (start, stop, provision)
• Configuration management endpoints

Comparison with Other Topologies

Troubleshooting

Problem: Cannot perform switchover in Active-Passive mode

Solution: Active-Passive mode disables failover/switchover by design. Either:

• Disable Active-Passive mode: replication-active-passive = false
• Manually promote new master outside replication-manager

Problem: replication-manager not detecting replication

Solution: Active-Passive mode ignores replication status. If replication monitoring is needed:

• Switch to appropriate topology (master-slave, multi-master, etc.)
• Remove replication-active-passive = true from configuration

Problem: Multiple servers all showing as "Unconnected"

Solution: In Active-Passive with multiple servers, this is expected. Only the first server is marked as master. To change active server:

• Use API: PUT /api/clusters/{cluster}/servers/{server}/actions/set-master
• Or configure db-servers-prefered-master to specify desired active server

Best Practices

Monitoring without disruption:

• Use Active-Passive mode when replication-manager should observe but not modify infrastructure
• Ideal for migration periods or auditing existing setups

Backup automation:

• Active-Passive mode fully supports backup scheduling and management
• Use restic integration for point-in-time recovery

Schema tracking:

• Enable monitoring-schema-change to detect unplanned schema modifications
• Configure monitoring-schema-change-script for alerting

Metrics collection:

• Active-Passive mode continues collecting all performance metrics
• Use Graphite or Prometheus endpoints for visualization in Grafana

Security:

• Use dedicated monitoring user with minimal privileges
• No REPLICATION privilege required in Active-Passive mode
• Recommended grants: PROCESS, REPLICATION CLIENT, SELECT