default_branch: "" => "<computed>"

description: "" => "my terraform repo"

etag: "" => "<computed>"

full_name: "" => "<computed>"

git_clone_url: "" => "<computed>"

html_url: "" => "<computed>"

http_clone_url: "" => "<computed>"

name: "" => "terraform-repo2"

ssh_clone_url: "" => "<computed>"

svn_url: "" => "<computed>"

github_repository.terraform_repo: Creation complete after 5s (ID: terraform-repo2)

Apply complete! Resources: 1 added, 0 changed, 1 destroyed.

For reasons of clarity, I created a big security hole – I put the token in the configuration file, and therefore in the repository, and now anyone who can access it can delete all repositories. Terraform provides several ways to set variables besides the one used. I'll just recreate the token and override it with the one passed on the command line:

(agile-aleph-203917) $ rm variables.tf

(agile-aleph-203917) $ sed -i 's / terraform-repo2 / terraform-repo3 /' main.tf

./terraform apply -var = "github_token = f7602b82e02efcbae7fc915c16eeee518280cf2a"

Building infrastructure in GCP with Terraform

Each cloud has its own set of services, its own APIs for them. To simplify the transition from one cloud for both employees in terms of learning and rewriting, there are universal libraries that implement the Facade pattern. A facade is understood as a universal API that disrupts the features of the systems behind it.

One representative of the cloud API facades is KOPS. KOPS is a tool for deploying Kubernetes to GCP, AWS and Azure. KOPS is similar to Kubectl – it is a binary, it can create both commands and the YML config, has a similar syntax, but unlike Kubectl, it creates not a POD, but a cluster node. Another example is Terraform, which specializes in deployment by configuration to adhere to the IasC concept.

To create the infrastructure, we need a token, it is created in GCP for the service account to which access is issued. To do this, I went along the path: IAM and administration -> Service accounts -> Create a service account and upon creation I dropped the Owner role (full access for test purposes), created a key with the Create key button in JSON format and renamed the downloaded key to Key. JSON. To describe the infrastructure, I used the documentation www.terraform.io/docs/providers/google/index.html :

(agil7e-aleph-20391) $ cat main.tf

provider "google" {

credentials = "$ {file (" key.json ")}"

project = "agile-aleph-203917"

region = "us-central1"

}

resource "google_compute_instance" "terraform" {

name = "terraform"

machine_type = "n1-standard-1"

zone = "us-central1-a"

boot_disk {

initialize_params {

image = "debian-cloud / debian-9"

}

}

network_interface {

network = "default"

}

}

Let's check the user rights:

(agile-aleph-203917) $ gcloud auth list

Credentialed Accounts

ACTIVE ACCOUNT

* esschtolts@gmail.com

To set the active account, run:

$ gcloud config set account `ACCOUNT`

Let's select the project as the current one (you can create the current one by default):

$ gcloud config set project agil7e-aleph-20391;

(agil7e-aleph-20391) $ ./terraform init | grep success

Terraform has been successfully initialized!

Now let's create one instance in the WEB console, after copying the key to the key.json file in the Terraform directory:

machine_type: "" => "n1-standard-1"

metadata_fingerprint: "" => "<computed>"

name: "" => "terraform"

network_interface. #: "" => "1"

network_interface.0.address: "" => "<computed>"

network_interface.0.name: "" => "<computed>"

network_interface.0.network: "" => "default"

network_interface.0.network_ip: "" => "<computed>"

network_interface.0.network: "" => "default"

project: "" => "<computed>"

scheduling. #: "" => "<computed>"

self_link: "" => "<computed>"

tags_fingerprint: "" => "<computed>"

zone: "" => "us-central1-a"

google_compute_instance.terraform: Still creating … (10s elapsed)

google_compute_instance.terraform: Still creating … (20s elapsed)

google_compute_instance.terraform: Still creating … (30s elapsed)