Read replicas in GCP - Step-by-Step Execution

Process Flow - Read replicas

Primary Database

↓

Create Read Replica

↓

Replica Syncs Data

↓

Read Traffic Directed to Replica

↓

Replica Handles Read Queries

↓

Primary Handles Writes

The primary database creates a read replica that syncs data continuously. Read queries go to the replica, while writes stay on the primary.

Execution Sample

GCP

gcloud sql instances create my-replica \
  --master-instance-name=my-primary \
  --region=us-central1

This command creates a read replica named 'my-replica' for the primary instance 'my-primary' in the specified region.

Process Table

Step	Action	Result	State Change
1	Create primary database instance	Primary instance 'my-primary' created	Primary database ready
2	Run command to create read replica	Read replica 'my-replica' creation started	Replica instance initializing
3	Replica syncs data from primary	Data replication in progress	Replica data catching up
4	Replica sync completes	Replica data up-to-date	Replica ready to serve reads
5	Application sends read queries	Queries routed to replica	Replica handles read traffic
6	Application sends write queries	Writes sent to primary	Primary handles writes
7	Replica continues syncing changes	Replica stays updated	Continuous replication ongoing

💡 Read replica is fully operational and serving read traffic while primary handles writes.

Status Tracker

Variable	Start	After Step 2	After Step 4	After Step 7
Primary Instance	Not created	Created and running	Running	Running
Read Replica	Not created	Initializing	Ready and synced	Ready and continuously syncing
Data Sync Status	None	Starting sync	Synced	Continuous sync ongoing
Read Traffic	None	None	Starts routing to replica	Reads served by replica
Write Traffic	None	None	Handled by primary	Handled by primary

Key Moments - 3 Insights

Why does the read replica not handle write queries?

Because the read replica is designed only to serve read queries to reduce load on the primary. Writes must go to the primary to keep data consistent, as shown in execution_table steps 5 and 6.

What happens if the replica is not fully synced?

If the replica is not synced (step 3), it cannot serve accurate read queries. The system waits until sync completes (step 4) before routing reads to the replica.

Does the replica operate independently from the primary?

No, the replica continuously syncs data from the primary (step 7). It depends on the primary for updates and cannot accept writes.

Visual Quiz - 3 Questions

Look at the execution_table, at which step does the replica become ready to serve read queries?

A Step 2

B Step 4

C Step 5

D Step 7

💡 Hint

Check the 'Result' column for when the replica data is up-to-date and ready.

According to variable_tracker, what is the status of 'Data Sync Status' after step 7?

A None

B Synced

C Continuous sync ongoing

D Starting sync

💡 Hint

Look at the last column for 'Data Sync Status' in variable_tracker.

If the application sends a write query, where does it go according to execution_table?

A Primary database

B Both primary and replica

C Read replica

D Neither

💡 Hint

Refer to step 6 in execution_table under 'Action' and 'Result'.

Concept Snapshot

Read replicas copy data from a primary database to handle read queries.
Writes always go to the primary to keep data consistent.
Replica syncs continuously to stay updated.
This reduces load on the primary and improves read performance.
In GCP, create replicas with 'gcloud sql instances create' using --master-instance-name.

Full Transcript

Read replicas are copies of a primary database that handle read queries to reduce load on the primary. The primary database handles all write queries to maintain data consistency. When you create a read replica in GCP, it initializes and syncs data from the primary. Once synced, the replica serves read traffic while continuously syncing updates. This setup improves performance by distributing read load. The replica cannot accept writes and depends on the primary for data updates.