Database High Availability
summary
Concrete comparison of high-availability options for MariaDB and PostgreSQL, including tradeoffs between logical backups (pg_dump) and physical file-level/hot backups plus WAL/transaction-log replay to restore systems in minutes to an hour even after years of activity. Practical procedures for creating a replica from a hot backup, using IP failover or a load balancer for zero-downtime maintenance, and the operational differences between synchronous (ack-wait) and asynchronous replication across datacenters are covered, with emphasis on open-source tools and exact failure modes observed.
AI-written summary · about
chapters (6)
- 0:00Start
- 15:25Physical Backup
- 28:52High Availability
- 39:55How To Set Up a New Slave
- 44:24Questions
- 49:22What Happens with Uncommitted Transactions on a Crash
related talks
50mMoving from MSSQL to pgSQL – High Availability and Disaster Recovery
46mDeploying Maximum HA Architecture With Postgres
24mComparing the High Availability of MySQL with Modern Distributed SQL Databases
36mDesigning Layered High‑Availability Architectures for PostgreSQL on a Budget
50mEvolution Of PostgreSQL Replication and High Availability (upto ~ 99.999%)
47mPostgreSQL High Availability: The Considerations and the Candidates
29mEverything You Ever Wanted to Know About MariaDB Replication but Were Afraid to Ask
52mOverview of Solutions for PostgreSQL High Availability