- Terraform >= v0.12
- Node.js >= v12.16
- Java (Amazon Corretto, any OpenJDK distribution, or Oracle's JDK) >= 8
- Maven >= 3.5
Currently cdktf is distributed only on GitHub packages. Until it is available in Maven Central, some additional configuration is needed.
Specifically, add https://maven.pkg.github.com/hashicorp/terraform-cdk
as a repository and add GitHub credentials using a Personal Access Token.
For more details read through GitHub's Guide.
Install with Homebrew:
$ brew install cdktf
Or install with npm
(comes with Node.js):
npm install -g cdktf-cli
Learn more how to use the cdktf command-line interface here.
mkdir hello-terraform
cd hello-terraform
Initialize a new CDK for Terraform project
cdktf init --template="java" --local
This will initialize a brand new CDK for Terraform project in Java using an interactive command. While this template uses Maven, it is possible to use Gradle or other build tools.
Note: By supplying '--local' option you have chosen local storage mode for storing the state of your stack.
This means that your Terraform state file will be stored locally on disk in a file 'terraform.tfstate' in the root of your project.
We will now set up the project. Please enter the details for your project.
If you want to exit, press ^C.
Project Name: (default: 'hello-terraform')
Project Description: (default: 'A simple getting started project for cdktf.')
Also, this command installs the cdktf
library so that it can be used in the project.
You can now open up the Main.java
file to view your application code.
vim src/main/java/com/mycompany/app/Main.java
package com.mycompany.app;
import software.constructs.Construct;
import com.hashicorp.cdktf.App;
import com.hashicorp.cdktf.TerraformStack;
public class Main extends TerraformStack
{
public Main(final Construct scope, final String id) {
super(scope, id);
// define resources here
}
public static void main(String[] args) {
final App app = new App();
new Main(app, "hello-terraform");
app.synth();
}
}
Refer to the examples directory for additional examples.
Let's take a simple Java application that uses the CDK for Terraform package.
package com.mycompany.app;
import software.constructs.Construct;
import com.hashicorp.cdktf.App;
import com.hashicorp.cdktf.TerraformStack;
import imports.aws.AwsProvider;
import imports.aws.Instance;
public class Main extends TerraformStack
{
public Main(final Construct scope, final String id) {
super(scope, id);
AwsProvider.Builder.create(this, "Aws")
.region("us-east-1")
.build();
Instance.Builder.create(this, "hello")
.ami("ami-2757f631")
.instanceType("t2.micro")
.build();
}
public static void main(String[] args) {
final App app = new App();
new Main(app, "hello-terraform");
app.synth();
}
}
As you can see in the example above we want to deploy AWS resources, therefore we need to use the Terraform AWS Provider. Every Terraform Provider and Terraform Module can be used with the workflow described here.
You specify it in your local cdktf.json
file, like this:
{
"language": "java",
"app": "mvn -e -q compile exec:java",
"terraformProviders": [
// Terraform Providers with version constraint go here
"aws@~> 3.45"
],
"terraformModules": [
// Terraform Modules with version constraint go here
],
"context": {
"excludeStackIdFromLogicalIds": "true"
}
}
Next you need to run cdktf get
which generates provider-specific bindings.
To save some time you can also use pre-built providers, please visit the "Using Providers"-Guide for more on that topic.
When you are ready you can run the synthesize
command to generate Terraform JSON configuration for the application.
cdktf synth
Generated Terraform code in the output directory: cdktf.out
This command will generate a directory called cdktf.out
. This directory contains the Terraform JSON configuration for
the application. The Terraform config for your stack hello-terraform
will be in cdktf.out/stacks/hello-terraform
.
cd cdktf.out/stacks/hello-terraform
Terraform AWS provider and instance expressed as Terraform JSON configuration.
cat cdk.tf.json
{
"terraform": {
"required_providers": {
"aws": "~> 2.0"
}
},
"provider": {
"aws": [
{
"region": "us-east-1"
}
]
},
"resource": {
"aws_instance": {
"examplejavahello3532B955": {
"ami": "ami-2757f631",
"instance_type": "t2.micro"
}
}
}
}
Note: You can generate the Terraform JSON configuration while synthesizing the code by running
cdktf synth --json
.
Note: You can use Terraform commands like
terraform init
,terraform plan
, andterraform apply
with the generated Terraform JSON configuration (learn more here) or optionally continue to use the CDK for Terraform CLI for a first-class experience.
You can now deploy your CDK for Terraform application using the cdktf deploy
command.
cdktf deploy
This command will ask for confirmation on a generated diff and then deploy the application.
Stack: helloterraform
Resources
+ AWS_INSTANCE hello aws_instance.helloterraform_hello_A790281A
Diff: 1 to create, 0 to update, 0 to delete.
Do you want to continue (Y/n)? y
Deployed application
Deploying Stack: helloterraform
Resources
✔ AWS_INSTANCE hello aws_instance.helloterraform_hello_A790281A
Summary: 1 created, 0 updated, 0 destroyed.
The cdktf deploy
command runs a terraform apply
in the background. If you are using local storage mode then it creates a terraform.tfstate
file in the root of the project.
You can destroy the application by running cdktf destroy
.
cdktf destroy
This command will ask for confirmation on a generated diff and then destroy the application if the user confirms that they want to continue with the destroy operation.
Stack: helloterraform
Resources
- AWS_INSTANCE hello aws_instance.helloterraform_hello_A790281A
Diff: 0 to create, 0 to update, 1 to delete.
Do you want to continue (Y/n)?
Destroyed application
Destroying Stack: helloterraform
Resources
✔ AWS_INSTANCE hello aws_instance.helloterraform_hello_A790281A
Summary: 1 destroyed.