Best Practices for MySQL High Availability in 2022
summary
Concrete recommendations and measurable tradeoffs: prefer ProxySQL for rich routing and multiplexing, prefer InnoDB Cluster (MySQL 8.0) or Galera variants (Percona XtraDB Cluster, MariaDB Galera) for virtually synchronous multi-master, require three-node quorum, expect multi-AZ/DRBD setups to cut worst-case throughput by up to ~60%, use semi-synchronous replication (used at Facebook since 2014) to avoid data loss, enable sync_binlog=1 and innodb_flush_log_at_trx_commit=1, and use the MySQL clone plugin or XtraBackup/MariaBackup for fast provisioning and SST.
AI-written summary · about
chapters (9)
- 0:00Start
- 3:42Modern Versions of Mysql and Mariadb
- 9:18Group Commit
- 22:35Using Multi-Threaded Slaves
- 23:28Time Delayed Replication
- 29:19Parallel Replication
- 30:11Gtid Migration
- 30:42Schema Parallel Replication
- 50:47Conclusion
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
53mMySQL DevOps
46mDatabase Reliability Engineering for MySQL
27mDistributed MySQL Architectures: Past, Present and Future
14mThe State of the Art of MySQL Replication
49mMySQL Replication — Advanced Features
