Terraform 0.2: What's New, Release History & EOL Status

What Is New in Terraform 0.2

Terraform 0.2 introduces foundational features that establish the core patterns for infrastructure as code. This release focuses on making remote state management a first-class concept and introduces the primitives for defining and connecting infrastructure components. These changes are the bedrock upon which all future Terraform functionality is built.

How does remote state work in Terraform 0.2?

Remote state is configured using the new terraform remote command, which stores the terraform.tfstate file on a shared server instead of locally. This is a game-changer for team collaboration because it prevents state file conflicts and provides a single source of truth for your infrastructure's current state.

To reference another state file, you use the terraform_remote_state data source. This lets you pull output values from a different Terraform configuration, enabling you to connect infrastructure stacks. In practice, this is how you make your network configuration aware of your compute instances.

# Configuring remote state
terraform remote config -backend=consul -backend-config="address=consul.example.com:8500"

# Referencing remote state in a data source
data "terraform_remote_state" "network" {
    backend = "consul"
    config {
        address = "consul.example.com:8500"
        path = "terraform/network"
    }
}

# Using an output from the remote state
resource "aws_instance" "app" {
    subnet_id = "${data.terraform_remote_state.network.output.subnet_id}"
}

What are the new push and pull commands for?

The terraform push and pull commands are the manual interface for managing your remote state. terraform push uploads your local state file to the configured remote backend, while pull downloads the remote state to your local machine. This gives you direct control over state synchronization.

You'll use pull to ensure you're working with the latest state before making changes, and push to publish your state updates after a successful apply. This manual step was necessary in early versions before automatic state locking and syncing were implemented.

How have variables and outputs changed?

Variables and outputs are now formally defined in your configuration using dedicated blocks. This moves them from being simple string replacements to being managed, typed values within Terraform's workflow. You define variables with variable blocks and declare their values in .tfvars files.

Outputs now have their own output blocks, making them explicit parts of your configuration that are persisted in the state file. This matters because it allows the terraform_remote_state data source to read these values, enabling cross-stack dependencies.

# Defining a variable
variable "instance_count" {
    description = "Number of instances to create"
    default = 2
}

# Defining an output
output "instance_ips" {
    value = ["${aws_instance.app.*.public_ip}"]
}