setup terraform

terraform-install | yandex-cli | terraform

setupsetup

mkdir -p ~/dev/cloud-terraform/
cd ~/dev/cloud-terraform/
vi providers.tf

terraform {
  required_version = ">= 1.0.0"

  required_providers {
yandex = {
  source = "yandex-cloud/yandex"
}
  }
}

prepare compute node creationprepare compute node creation

–either– find your favorite image branch/family from the cli – BEWARE you need to increase output limit AFTER specifying folder-id

yc compute image list --format json --folder-id standard-images --limit 10000 | \
    jq -r '.[] | select( .family == "debian-11-oslogin" )' | \
    jq -r '.name + "," + .id + "," + .status' | sort -V | tail -5

yc compute image list --format json --folder-id standard-images --limit 10000 | \
    jq -r '.[] | select( .family == "debian-12" )' | \
    jq -r '.name + "," + .id + "," + .status ' | sort -V | tail -5

–or– just select one from the web console

Create instance
Marketplace > search Debian

minimum hw resourcesminimum hw resources

https://yandex.cloud/en/docs/compute/concepts/vm-platforms

https://yandex.cloud/en/docs/compute/concepts/performance-levels

we want

standard-v4a

here’s the minimal setup

2 vCPU
1 GB RAM
garanteed 20%

and further minimum settings

SSD 5GB
preemptible

that is

# as of jan 2026
image_id=fd8miiisblcuktpjr6sc

# <FOLDER>-ru-central1-d
zone_subnet_id=...

# openssh format
pubkey="..."

cat > main.tf <<EOF
resource "yandex_compute_disk" "boot" {
  name     = "testvm"
  type     = "network-ssd"
  zone     = "ru-central1-d"
  size     = "5"
  image_id = "$image_id"
}

resource "yandex_compute_instance" "this" {
  name  = "testvm"
  platform_id = "standard-v4a"
  zone  = "ru-central1-d"

  resources {
    cores    = 2
    memory  = 1
    core_fraction = 20
  }

  boot_disk {
    disk_id = yandex_compute_disk.boot.id
  }

  # you need a public ip unless you have access to the internal subnet
  network_interface {
    subnet_id = "$zone_subnet_id"
    nat       = true
  }

  # username does not matter, as it depends on the image
  metadata = {
    ssh-keys       = "debian:$pubkey"
    serial-port-enable = 1
  }

  scheduling_policy {
    preemptible = true
  }

  allow_stopping_for_update = true
}
EOF

note

  # changing hostname re-defines everything - avoid that here even for the first shot
  #hostname    = "test"

moar (unused)

    # enable only once helper tools are installed
    enable-oslogin = false

ready to goready to go

rm -rf .terraform*
terraform init

yc config profile list
yc config profile activate test

export YC_TOKEN=`yc iam create-token`
export YC_CLOUD_ID=`yc config get cloud-id`
export YC_FOLDER_ID=`yc config get folder-id`

echo $YC_TOKEN
echo $YC_CLOUD_ID
echo $YC_FOLDER_ID

terraform plan
terraform apply

importimport

terraform import -no-color

resourcesresources

https://yandex.cloud/en/docs/tutorials/infrastructure-management/terraform-quickstart#cli_1

https://yandex.cloud/en/docs/tutorials/infrastructure-management/terraform-modules

https://yandex.cloud/en/docs/tutorials/infrastructure-management/terraform-data-sources

https://terraform-provider.yandexcloud.net/Resources/compute_instance

https://terraform-provider.yandexcloud.net/Resources/organizationmanager_os_login_settings

https://github.com/yandex-cloud/terraform-provider-yandex

https://registry.terraform.io/providers/yandex-cloud/yandex/latest

https://spacelift.io/blog/importing-exisiting-infrastructure-into-terraform