data-replication

name	data-replication
description	Design data replication strategies for high availability, read scaling, and disaster recovery. Outputs replication topology, consistency model, failover procedures, and lag monitoring setup.
argument-hint	["database type","consistency requirements","read/write ratio","RTO/RPO targets","geographic distribution"]
allowed-tools	Read, Write

Data Replication

Data replication maintains copies of data across multiple nodes to achieve read scaling, high availability, and disaster recovery. The fundamental tension is between consistency (all replicas have the same data) and availability (reads and writes succeed even when nodes fail).

Replication Topologies

1. Single-leader (Primary/Replica)
   Primary: all writes
   Replicas: read scale + failover
   Consistency: eventual (async) or synchronous (1+ replica)
   
2. Multi-leader
   Multiple nodes accept writes
   Conflict resolution required
   Use: multi-datacenter, offline-capable clients
   
3. Leaderless (Dynamo-style)
   Any node accepts reads and writes
   Quorum: W + R > N for consistency
   Use: Cassandra, DynamoDB, Riak

PostgreSQL Streaming Replication

-- Primary: postgresql.conf
wal_level = replica
max_wal_senders = 5
max_replication_slots = 5
wal_keep_size  GB
synchronous_commit             
synchronous_standby_names      


host replication replicator replica1_ip scramsha
host replication replicator replica2_ip scramsha


  replicator REPLICATION LOGIN ENCRYPTED PASSWORD ;


primary_conninfo  
primary_slot_name  
hot_standby                 
hot_standby_feedback        


 client_addr, state, sent_lsn, write_lsn, flush_lsn, replay_lsn,
       (sent_lsn  replay_lsn)  replication_lag_bytes,
       (epoch  (now()  replay_lag))  lag_seconds
 pg_stat_replication;


 pg_create_physical_replication_slot();
 slot_name, active, pg_wal_lsn_diff(pg_current_wal_lsn(), restart_lsn)  lag_bytes
 pg_replication_slots;