AWS Aurora for PostgreSQL

Aurora for PostgreSQL is PostgreSQL-compatible. The setup is identical to RDS with one critical difference: you must connect to the writer endpoint, not a reader endpoint or an instance endpoint.

Constraints

All RDS constraints apply, plus:

Connect to the cluster writer endpoint. Aurora readers are read-only replicas. pg_auto_mv writes to pgauto_mv.mv_state, pgauto_mv.mv_events, and pgauto_mv.mv_refresh_log during normal operation. All DML must target the writer.

Your cluster endpoints:

Endpoint type	Format	Use
Writer (use this)	my-cluster.cluster-abc123.us-east-1.rds.amazonaws.com	All pg_auto_mv installation and operation
Reader (do not use)	my-cluster.cluster-ro-abc123.us-east-1.rds.amazonaws.com	Application reads only

Step 1 — Create the pgauto_mv Schema

psql -h my-cluster.cluster-abc123.us-east-1.rds.amazonaws.com \
     -U postgres \
     -d your_database \
     -c "CREATE SCHEMA pgauto_mv;"

Step 2 — Deploy the pg_auto_mv Objects

psql -h my-cluster.cluster-abc123.us-east-1.rds.amazonaws.com \
     -U postgres \
     -d your_database \
     -f sql/pg_auto_mv--1.0.sql \
     -f sql/pg_auto_mv--1.0--1.1.sql

Step 3 — Verify the Installation

SELECT channel, action
FROM pgrelay.actions
WHERE channel LIKE 'pg_auto_mv.%'
ORDER BY channel;

Failover Behaviour

During an Aurora failover, the pg_relay Processor disconnects and reconnects to the new primary via the writer endpoint. Any pg_auto_mv events that were enqueued in pgrelay.queue before the failover will be processed when the Processor reconnects — no events are lost. The materialised view state (pgauto_mv.mv_state) is preserved across failover by Aurora's standard replication.

---