Read replicas in GCP - Mini Project: Build & Apply

Setting Up Read Replicas in Google Cloud SQL

📖 Scenario: You are managing a Google Cloud SQL database for a growing web application. To improve read performance and reduce load on the primary database, you want to create read replicas.This project will guide you through creating a Cloud SQL instance, configuring a read replica, and setting the necessary parameters.

🎯 Goal: Build a Google Cloud SQL primary instance and a read replica instance using Terraform configuration files.This setup will help distribute read traffic and improve database scalability.

📋 What You'll Learn

Create a primary Cloud SQL instance with MySQL version 8.0

Define a variable for the instance region

Create a read replica instance linked to the primary instance

Configure the read replica with appropriate settings

💡 Why This Matters

🌍 Real World

Read replicas help distribute database read traffic, improving performance and scalability for applications with many users.

💼 Career

Cloud architects and DevOps engineers often configure read replicas to optimize database workloads and ensure high availability.

1

Create the primary Cloud SQL instance resource

Create a Terraform resource called google_sql_database_instance named primary_instance with name = "primary-db-instance", database_version = "MYSQL_8_0", and region = "us-central1". Use the settings block with tier = "db-f1-micro".

GCP

# Your code here to create the primary Cloud SQL instance resource

Hint

Use the google_sql_database_instance resource with the exact name primary_instance. Set the name, database_version, and region as specified. Inside settings, set the tier to db-f1-micro.

2

Add a variable for the instance region

Create a Terraform variable called instance_region with default = "us-central1".

GCP

resource "google_sql_database_instance" "primary_instance" {
  name             = "primary-db-instance"
  database_version = "MYSQL_8_0"
  region           = "us-central1"

  settings {
    tier = "db-f1-micro"
  }
}

# Your code here to add the variable for instance_region

Hint

Define a variable named instance_region with the default value "us-central1". Then update the region in the primary instance resource to use var.instance_region.

3

Create the read replica instance resource

Create a Terraform resource called google_sql_database_instance named read_replica with name = "read-replica-instance", region = var.instance_region, and database_version = "MYSQL_8_0". Add a replica_configuration block with replica_name = "read-replica-1" and primary_instance_name = google_sql_database_instance.primary_instance.name. Use the settings block with tier = "db-f1-micro".

GCP

variable "instance_region" {
  default = "us-central1"
}

resource "google_sql_database_instance" "primary_instance" {
  name             = "primary-db-instance"
  database_version = "MYSQL_8_0"
  region           = var.instance_region

  settings {
    tier = "db-f1-micro"
  }
}

# Your code here to create the read replica resource

Hint

Create a new resource named read_replica of type google_sql_database_instance. Set the name, region, and database_version. Inside replica_configuration, set replica_name and link to the primary instance using primary_instance_name. Add settings with the tier.

4

Enable backup and maintenance window on the primary instance

Add a backup_configuration block to google_sql_database_instance.primary_instance with enabled = true. Also add a maintenance_window block with day = 7, hour = 3, and update_track = "stable".

GCP

variable "instance_region" {
  default = "us-central1"
}

resource "google_sql_database_instance" "primary_instance" {
  name             = "primary-db-instance"
  database_version = "MYSQL_8_0"
  region           = var.instance_region

  settings {
    tier = "db-f1-micro"
  }

  # Your code here to add backup_configuration and maintenance_window
}

resource "google_sql_database_instance" "read_replica" {
  name             = "read-replica-instance"
  region           = var.instance_region
  database_version = "MYSQL_8_0"

  replica_configuration {
    replica_name          = "read-replica-1"
    primary_instance_name = google_sql_database_instance.primary_instance.name
  }

  settings {
    tier = "db-f1-micro"
  }
}

Hint

Inside the primary_instance resource, add a backup_configuration block with enabled = true. Also add a maintenance_window block with the specified day, hour, and update_track values.

Practice

1.

What is the main purpose of a read replica in Google Cloud SQL?

easy

A. To copy the main database and handle read requests to reduce load

B. To store backups of the database for disaster recovery

C. To serve as a primary database for write operations

D. To encrypt data stored in the database

Solution

1. Step 1: Understand the role of read replicas

   Read replicas copy data from the main database and serve read requests to reduce load on the primary instance.

2. Step 2: Differentiate from other database roles

   Backups and encryption are different features; read replicas do not handle writes.

3. Final Answer:

   To copy the main database and handle read requests to reduce load -> Option A

4. Quick Check:

   Read replica purpose = reduce load by handling reads [OK]

Hint: Read replicas handle reads to reduce main database load [OK]

Common Mistakes:

• Confusing read replicas with backups
• Thinking read replicas handle writes
• Assuming read replicas encrypt data

2.

Which gcloud command correctly creates a read replica named replica-instance of a master instance called master-instance?

easy

A. gcloud sql instances create replica-instance --master-instance-name=master-instance

B. gcloud sql instances create replica-instance --replica-of=master-instance

C. gcloud sql replicas create replica-instance --master=master-instance

D. gcloud sql create instance replica-instance --master=master-instance

Solution

1. Step 1: Recall the correct command syntax

   The correct command to create a read replica uses gcloud sql instances create with the flag --master-instance-name.

2. Step 2: Verify option correctness

   gcloud sql instances create replica-instance --master-instance-name=master-instance matches the correct syntax exactly; other options use incorrect flags or commands.

3. Final Answer:

   gcloud sql instances create replica-instance --master-instance-name=master-instance -> Option A

4. Quick Check:

   Correct flag for read replica creation = --master-instance-name [OK]

Hint: Use --master-instance-name flag with gcloud sql instances create [OK]

Common Mistakes:

• Using incorrect flags like --replica-of
• Using wrong gcloud subcommands
• Mixing up command order or syntax

3.

Consider this command output snippet after creating a read replica:

NAME              REGION       MASTER_INSTANCE
replica-instance  us-central1  master-instance

What does this output indicate?

medium

A. The master-instance is a read replica of replica-instance

B. The replica-instance is the primary database in us-central1 region

C. The replica-instance is a read replica of master-instance in us-central1 region

D. The replica-instance is a backup stored in us-central1

Solution

1. Step 1: Analyze the output columns

   The output shows the replica name, its region, and the master instance it replicates.

2. Step 2: Interpret the relationship

   Since replica-instance lists master-instance as its master, it is a read replica of that master in the specified region.

3. Final Answer:

   The replica-instance is a read replica of master-instance in us-central1 region -> Option C

4. Quick Check:

   Replica points to master-instance = read replica relationship [OK]

Hint: Master instance listed means replica points to it for reads [OK]

Common Mistakes:

• Confusing replica as primary
• Thinking master is replica
• Assuming backup instead of replica

4.

You tried to create a read replica with this command:

gcloud sql instances create replica1 --master=master1

But it failed. What is the likely cause?

medium

A. You must create the replica before the master instance

B. The replica name cannot contain numbers

C. The master instance must be in a different region

D. The flag --master is incorrect; should use --master-instance-name

Solution

1. Step 1: Check the command flags

   The correct flag to specify the master instance is --master-instance-name, not --master.

2. Step 2: Validate other options

   Replica names can contain numbers; master and replica can be in same region; master must exist before replica.

3. Final Answer:

   The flag --master is incorrect; should use --master-instance-name -> Option D

4. Quick Check:

   Correct flag for master instance = --master-instance-name [OK]

Hint: Use exact flag --master-instance-name to link replica to master [OK]

Common Mistakes:

• Using wrong flag --master
• Thinking replica name rules are stricter
• Assuming region mismatch causes failure

5.

You have a master instance handling both reads and writes. To improve read performance globally, you want to create two read replicas in different regions. Which approach is best?

1. Create two read replicas using gcloud sql instances create with --master-instance-name for each region.
2. Configure your application to send read queries to replicas and write queries to the master.
3. Ensure replicas are in regions close to your users for low latency.

What is the correct reasoning behind this approach?

hard

A. Read replicas automatically handle writes to reduce master load

B. Read replicas reduce load on master and improve read latency by being regionally closer to users

C. Creating replicas in multiple regions duplicates writes to all replicas instantly

D. Replicas replace the master for all database operations

Solution

1. Step 1: Understand read replica benefits

   Read replicas copy data from the master and serve read requests, reducing load on the master.

2. Step 2: Consider regional placement and query routing

   Placing replicas near users lowers latency; routing reads to replicas and writes to master optimizes performance.

3. Step 3: Clarify write behavior

   Replicas do not handle writes; writes go only to the master and replicate asynchronously.

4. Final Answer:

   Read replicas reduce load on master and improve read latency by being regionally closer to users -> Option B

5. Quick Check:

   Replicas improve read speed and reduce master load [OK]

Hint: Replicas serve reads near users; writes stay on master [OK]

Common Mistakes:

• Assuming replicas handle writes
• Thinking replicas instantly sync writes synchronously
• Believing replicas replace the master