Remote state data source for cross-project in Terraform - Mini Project: Build & Apply

Remote state data source for cross-project

📖 Scenario: You are managing infrastructure for two separate projects in Terraform. Project A creates a network, and Project B needs to use the network's ID from Project A to create a virtual machine inside that network.

🎯 Goal: Learn how to use Terraform's remote state data source to access outputs from one project in another project.

📋 What You'll Learn

Create a Terraform backend configuration for Project A's state

Output the network ID from Project A

Configure Project B to read Project A's remote state

Use the network ID from Project A in Project B's resource

💡 Why This Matters

🌍 Real World

Many organizations split infrastructure into multiple Terraform projects for separation of concerns. Sharing outputs via remote state allows projects to connect resources safely.

💼 Career

Understanding remote state data sources is essential for Terraform users working in teams or managing complex infrastructure across multiple projects.

1

Create Project A's network and output

In Project A's Terraform configuration, create a resource called aws_vpc named main_network with cidr_block set to "10.0.0.0/16". Then create an output called network_id that outputs aws_vpc.main_network.id.

Terraform

# Your code here to create aws_vpc main_network and output network_id

Hint

Use resource "aws_vpc" "main_network" to create the network and output "network_id" to expose its ID.

2

Configure backend for Project A state storage

Add a terraform block with a backend configuration using s3 backend. Set bucket to "project-a-terraform-state", key to "network/terraform.tfstate", and region to "us-east-1".

Terraform

resource "aws_vpc" "main_network" {
  cidr_block = "10.0.0.0/16"
}

output "network_id" {
  value = aws_vpc.main_network.id
}

# Add terraform backend configuration here

Hint

Use the terraform block with backend "s3" and set the bucket, key, and region exactly as given.

3

Configure Project B to read Project A's remote state

In Project B's Terraform configuration, create a data block of type terraform_remote_state named project_a. Set the backend to s3 with bucket as "project-a-terraform-state", key as "network/terraform.tfstate", and region as "us-east-1".

Terraform

# Your code here to create terraform_remote_state data source named project_a

Hint

Use data "terraform_remote_state" "project_a" with backend = "s3" and the correct config values.

4

Use Project A's network ID in Project B resource

Create a resource aws_instance named vm in Project B. Set ami to "ami-12345678", instance_type to "t2.micro", and subnet_id to data.terraform_remote_state.project_a.outputs.network_id.

Terraform

data "terraform_remote_state" "project_a" {
  backend = "s3"
  config = {
    bucket = "project-a-terraform-state"
    key    = "network/terraform.tfstate"
    region = "us-east-1"
  }
}

# Your code here to create aws_instance vm using network_id from project_a outputs

Hint

Create resource "aws_instance" "vm" and set subnet_id to the remote state output.

Practice

1. What is the main purpose of using a terraform_remote_state data source in Terraform?

easy

A. To store Terraform state files locally

B. To access outputs from another Terraform project's state

C. To create new resources in a different cloud provider

D. To encrypt Terraform state files automatically

Solution

1. Step 1: Understand remote state data source role

   The terraform_remote_state data source allows one Terraform configuration to read outputs from another configuration's state file.

2. Step 2: Differentiate from other options

   It does not store state locally, create new resources, or encrypt state automatically; it only reads existing state outputs.

3. Final Answer:

   To access outputs from another Terraform project's state -> Option B

4. Quick Check:

   Remote state data source = Access outputs [OK]

Hint: Remote state data source reads outputs from other projects [OK]

Common Mistakes:

• Confusing remote state with local state storage
• Thinking it creates resources instead of reading state
• Assuming it encrypts state automatically

2. Which of the following is the correct syntax to define a terraform_remote_state data source for a backend stored in an S3 bucket named my-terraform-state?

easy

A. data "terraform_remote_state" "example" { backend = "local" config = { path = "my-terraform-state" } }

B. resource "terraform_remote_state" "example" { backend = "s3" bucket = "my-terraform-state" }

C. terraform_remote_state "example" { backend = "s3" bucket = "my-terraform-state" }

D. data "terraform_remote_state" "example" { backend = "s3" config = { bucket = "my-terraform-state" key = "state.tfstate" region = "us-east-1" } }

Solution

1. Step 1: Identify correct resource type and syntax

   The terraform_remote_state must be declared as a data block, not a resource.

2. Step 2: Check backend and config structure

   For S3 backend, the config requires bucket, key, and region inside a config map.

3. Final Answer:

   data "terraform_remote_state" "example" { backend = "s3" config = { bucket = "my-terraform-state" key = "state.tfstate" region = "us-east-1" } } -> Option D

4. Quick Check:

   Correct syntax = data "terraform_remote_state" "example" { backend = "s3" config = { bucket = "my-terraform-state" key = "state.tfstate" region = "us-east-1" } } [OK]

Hint: Use data block with backend and config map for remote state [OK]

Common Mistakes:

• Using resource instead of data block
• Missing required config keys like key or region
• Using wrong backend type like local for S3

3. Given this Terraform snippet accessing remote state:

data "terraform_remote_state" "network" {
  backend = "gcs"
  config = {
    bucket = "tf-state-bucket"
    prefix = "network"
  }
}

output "vpc_id" {
  value = data.terraform_remote_state.network.outputs.vpc_id
}

What will output.vpc_id contain?

medium

A. The entire remote state file content as a string

B. An error because prefix is not a valid config key for GCS backend

C. The VPC ID output from the remote state stored in the GCS bucket under prefix 'network'

D. Null because outputs cannot be accessed from remote state

Solution

1. Step 1: Understand remote state data source usage

   The data source reads the remote state from the GCS bucket with the given prefix, making outputs available.

2. Step 2: Confirm output access

   The output vpc_id is accessed correctly via data.terraform_remote_state.network.outputs.vpc_id, so it returns the VPC ID value.

3. Final Answer:

   The VPC ID output from the remote state stored in the GCS bucket under prefix 'network' -> Option C

4. Quick Check:

   Remote output access = VPC ID value [OK]

Hint: Remote state outputs accessed via data.<name>.outputs.<key> [OK]

Common Mistakes:

• Confusing prefix usage for GCS backend (it is valid)
• Expecting entire state file instead of outputs
• Assuming outputs cannot be read remotely

4. You have this Terraform remote state data source:

data "terraform_remote_state" "app" {
  backend = "azurerm"
  config = {
    resource_group_name = "rg-state"
    storage_account_name = "stterraform"
    container_name = "tfstate"
    key = "app.terraform.tfstate"
  }
}

When running terraform plan, you get an error: Failed to load remote state. What is the most likely cause?

medium

A. Incorrect or missing permissions to access the Azure storage account

B. The backend type should be s3 instead of azurerm

C. The key parameter is not supported in azurerm backend

D. Terraform remote state data source cannot be used with Azure

Solution

1. Step 1: Verify backend and config correctness

   The backend azurerm with given config keys is valid for Azure Blob Storage remote state.

2. Step 2: Identify common causes of load failure

   Most common cause is missing or incorrect permissions to access the storage account or container.

3. Final Answer:

   Incorrect or missing permissions to access the Azure storage account -> Option A

4. Quick Check:

   Access permissions issue = Load failure [OK]

Hint: Check storage permissions if remote state load fails [OK]

Common Mistakes:

• Assuming backend type is wrong when it is correct
• Thinking key parameter is unsupported in azurerm backend
• Believing remote state cannot be used with Azure

5. You manage two Terraform projects: network and app. The network project stores its state remotely in an S3 bucket with key network/terraform.tfstate. You want the app project to use the VPC ID output from network. Which configuration correctly sets up the remote state data source in app to access network outputs securely and follows best practices?

hard

A. data "terraform_remote_state" "network" {
backend = "s3"
config = {
bucket = "my-tf-state-bucket"
key = "network/terraform.tfstate"
region = "us-west-2"
}
}

B. data "terraform_remote_state" "network" {
backend = "s3"
config = {
bucket = "my-tf-state-bucket"
key = "app/terraform.tfstate"
region = "us-west-2"
}
}

C. data "terraform_remote_state" "network" {
backend = "local"
config = {
path = "../network/terraform.tfstate"
}
}

D. resource "terraform_remote_state" "network" {
backend = "s3"
config = {
bucket = "my-tf-state-bucket"
key = "network/terraform.tfstate"
region = "us-west-2"
}
}

Solution

1. Step 1: Confirm correct backend and key for remote state

   The remote state is stored in S3 bucket with key network/terraform.tfstate, so the data source must match this.

2. Step 2: Ensure data source type and security best practices

   Use a data block (not resource) with backend = "s3", specify region, and avoid incorrect keys.

3. Final Answer:

   Data source with backend s3, correct bucket/key, and region -> Option A

4. Quick Check:

   Correct backend and secure config = data "terraform_remote_state" "network" { backend = "s3" config = { bucket = "my-tf-state-bucket" key = "network/terraform.tfstate" region = "us-west-2" } } [OK]

Hint: Match backend/key exactly and use data block [OK]

Common Mistakes:

• Using wrong key path for remote state
• Using resource block instead of data block
• Using local backend instead of remote S3
• Omitting region config