How Safe is Asynchronous Master-Master Setup?
summary
Claims that built-in asynchronous active-active replication is unusable for production get challenged with concrete fixes: thousands of shops use asynchronous source-source replication with application routing, id namespaces and write-affinity to avoid conflicts. The talk shows specific failures—auto-increment collisions, replica lag causing lost updates, and split‑brain—and compares deployments that beat Galera/InnoDB Cluster for latency and partition tolerance, plus tested mitigation patterns and real postmortems that explain when the setup actually works.
AI-written summary · about
chapters (6)
- 0:00Start
- 10:44Fully Synchronous Cluster Node 1
- 12:21Synchronization Penalty Speed of the slowest member
- 13:21GEO Replication is Painful • Synchronization takes too much time .Nodes often disconnect
- 16:10Databases and Tables • Each server writes to its own set of tables
- 19:16More information MySQL User Reference Manual Why MySQL Replication Fails, and How... MySQL High Availability
related talks
35mMySQL HA, Recovery and Load Balancing
24mComparing the High Availability of MySQL with Modern Distributed SQL Databases
32mBest Practice in Design and Implementing MySQL Geographic Distributed, High Availability Solutions
41mAsynchronous Replication Without Censorship
52mBest Practices for MySQL High Availability in 2022
43mMySQL Failover for Asynchronous replication with Orchestrator, ProxySQL and Consul
14mThe State of the Art of MySQL Replication
49mMySQL Replication — Advanced Features