Comprendre les concepts de la réplication

Imagine you have a super important book. What if you lost it? You'd want a copy, right?

Database replication is similar! It's the process of creating and maintaining multiple copies of your database.

This ensures your data is safe and always available, even if something goes wrong with the original.

One of the biggest reasons for replication is High Availability.

• If your main database server (the original) crashes, you don't want your applications to stop working.
• With replication, another copy can quickly take over.
• This minimizes downtime and keeps your services running smoothly.

Think about a popular website. Thousands of users might be reading data at the same time.

If all these requests hit a single database, it can slow down significantly.

Replication allows you to distribute read queries across multiple copies, sharing the load and improving performance for everyone.

Disasters happen: hardware failures, natural events, or even human error.

Having your database replicated to a different physical location or server is crucial for disaster recovery.

If the primary data center goes offline, you can switch to a replicated copy, saving your business from data loss and extended outages.

The most common replication setup is the Primary-Standby (often called Master-Slave) architecture.

• The Primary server is the original database where all data changes (writes) occur.
• The Standby server(s) are the copies that receive changes from the primary.
• Standbys are typically read-only, serving as backups or for read scaling.

The Primary server is the heart of your replicated system. It's responsible for:

• Handling all incoming write operations (INSERT, UPDATE, DELETE).
• Generating the Write-Ahead Log (WAL), which records every change.
• Sending these WAL records to its Standby servers.

Think of it as the original author who writes new chapters and sends them to copy editors.

The Standby server's main purpose is to mirror the Primary. Its tasks include:

• Receiving WAL records from the Primary.
• Applying these WAL records to its own database, keeping it synchronized.
• Serving read-only queries, reducing load on the Primary.

It's like the copy editor who updates their copy with the new chapters from the author.

At the core of PostgreSQL replication is the Write-Ahead Log (WAL).

WAL is a sequence of records describing every change made to the database data files.

Before any actual data is written to disk, its change is recorded in the WAL. This ensures data integrity and durability.

WAL is essential for replication because:

• The Primary server streams its WAL records to the Standby server(s).
• The Standby server then "replays" these WAL records.
• By replaying the WAL, the Standby reconstructs the exact state of the Primary database.

This efficient mechanism ensures that Standbys are consistent copies of the Primary.

Replication offers many benefits. Which of these is NOT a primary reason to implement database replication?

You've just learned the fundamental concepts of database replication!

• Replication creates copies of your database for high availability, read scaling, and disaster recovery.
• The common setup involves a Primary server and one or more Standby servers.
• The Write-Ahead Log (WAL) is crucial, as it records all changes and is streamed to Standbys to keep them synchronized.

Next, we'll dive into setting up physical streaming replication!