A write conflict occurs when replicated changes cannot coexist with a primary-key, unique, or schema constraint. In bidirectional PostgreSQL logical replication, independently generated sequence values make this failure easy to trigger. section 5.5
Duplicate identity across nodes
After both nodes contain tickets with IDs 1–3, beta inserts another ticket while its sequence is still at 1. PostgreSQL rejects the replicated row with a duplicate-key error because alpha already used ID 1. Avoid reuse by synchronizing writes across the cluster or pre-allocating non-overlapping ranges, accepting the performance or administrative cost. section 5.5
Schema drift as a replication conflict
Renaming opened_at to logged_at on beta while alpha continues inserting leaves the target without the replicated column. The replication worker enters a crash/restart loop, so schema changes require coordinated orchestration. section 5.5
