Efficient MySQL Performance

Chapter 7. Replication Lag

Replication lag is the delay between the time when a write occurs on a source MySQL instance and the time when that write is applied on a replica MySQL instance. Replication lag is inherent to all database servers because replication across a network incurs network latency.

I’m glad that, as an engineer using MySQL, you don’t have to set up, configure, and maintain a MySQL replication topology because MySQL replication has become complex. Instead, this chapter investigates replication lag with respect to performance: what it is, why it happens, what risk it poses, and what you can do about it.

Simple Replication Won the Internet

Simple replication is one reason that MySQL became the most popular open source relational database server in the world. In the early 2000s, the internet was reemerging from the dot-com bubble of the ’90s and online companies were growing fast. Since replication is required for high availability and also used to scale out reads, simple replication in early versions of MySQL (v3.23 through v5.5) helped it win the internet in those heedless days. Early versions of MySQL used single-threaded statement-based replication (SBR): the source MySQL instance would log the SQL statements that it executed—yes, the actual SQL statements—and replica instances would simply re-execute those SQL statements. Replication doesn’t get simpler than that. Yes, it worked, and yes, it had problems and gotchas. But sometimes the simplest solution really ...