Develop Stack Template

Cluster.dev gives you freedom to modify existing templates or create your own. You can add inputs and outputs to already preset units, take the output of one unit and send it as an input to another, or write new units and add them to a template.

In our example we shall use the tmpl-development sample to create a K3s cluster on AWS. Then we shall modify the project stack template by adding new parameters to the units.

Workflow steps

1. Create a project following the steps described in Create Own Project section.

2. To start working with the stack template, cd into the template directory and open the template.yaml file: ./template/template.yaml.

   Our sample stack template contains 3 units. Now, let's elaborate on each of them and see how we can modify it.

3. The create-bucket unit uses a remote Terraform module to create an S3 bucket on AWS:

   name: create-bucket
   type: tfmodule
   providers: *provider_aws
   source: terraform-aws-modules/s3-bucket/aws
   version: "2.9.0"
   inputs:
     bucket: {{ .variables.bucket_name }}
     force_destroy: true
   

   We can modify the unit by adding more parameters in inputs. For example, let's add some tags using the insertYAML function:

   name: create-bucket
   type: tfmodule
   providers: *provider_aws
   source: terraform-aws-modules/s3-bucket/aws
   version: "2.9.0"
   inputs:
     bucket: {{ .variables.bucket_name }}
     force_destroy: true
     tags: {{ insertYAML .variables.tags }}
   

   Now we can see the tags in stack.yaml:

   name: cdev-tests-local
   template: ./template/
   kind: Stack
   backend: aws-backend
   variables:
     bucket_name: "tmpl-dev-test"
     region: {{ .project.variables.region }}
     organization: {{ .project.variables.organization }}
     name: "Developer"
     tags:
       tag1_name: "tag 1 value"
       tag2_name: "tag 2 value"
   

   To check the configuration, run cdev plan --tf-plan command. In the output you can see that Terraform will create a bucket with the defined tags. Run cdev apply -l debug to have the configuration applied.

4. The create-s3-object unit uses local Terraform module to get the bucket ID and save data inside the bucket. The Terraform module is stored in s3-file directory, main.tf file:

   name: create-s3-object
   type: tfmodule
   providers: *provider_aws
   source: ./s3-file/
   depends_on: this.create-bucket
   inputs:
     bucket_name: {{ remoteState "this.create-bucket.s3_bucket_id" }}
     data: |
       The data that will be saved in the S3 bucket after being processed by the template engine.
       Organization: {{ .variables.organization }}
       Name: {{ .variables.name }}
   

   The unit sends 2 parameters. The bucket_name is retrieved from the create-bucket unit by means of remoteState function. The data parameter uses templating to retrieve the Organization and Name variables from stack.yaml.

   Let's add to data input bucket_regional_domain_name variable to obtain the region-specific domain name of the bucket:

   name: create-s3-object
   type: tfmodule
   providers: *provider_aws
   source: ./s3-file/
   depends_on: this.create-bucket
   inputs:
     bucket_name: {{ remoteState "this.create-bucket.s3_bucket_id" }}
     data: |
       The data that will be saved in the s3 bucket after being processed by the template engine.
       Organization: {{ .variables.organization }}
       Name: {{ .variables.name }}
       Bucket regional domain name: {{ remoteState "this.create-bucket.s3_bucket_bucket_regional_domain_name" }}
   

   Check the configuration by running cdev plan command; apply it with cdev apply -l debug.

5. The print_outputs unit retrieves data from two other units by means of remoteState function: bucket_domain variable from create-bucket unit and s3_file_info from create-s3-object unit:

   name: print_outputs
   type: printer
   inputs:
     bucket_domain: {{ remoteState "this.create-bucket.s3_bucket_bucket_domain_name" }}
     s3_file_info: "To get file use: aws s3 cp {{ remoteState "this.create-s3-object.file_s3_url" }} ./my_file && cat my_file"
   

6. Having finished your work, run cdev destroy to eliminate the created resources.