Color boxes used in this tutorial
In this tutorial I'm using colored boxes according to the following conventions:
Useful summaries and takeaways.
Just random digressions and musings.
So, you might have heard about Terraform, but what exactly is Terraform?
Let's ask ChatGPT:
Hmm... I'm not sure what that means...
Let's look at another definition from https://kodekloud.com/playgrounds/playground-terraform:
Terraform is an infrastructure provisioning tool. It helps automate the process of creating and managing infrastructure using configuration files.
To provision infrastructure in Terraform, we follow three main steps: write, plan & apply.
Writing: In the writing stage, we define the infrastructure we want in a Terraform configuration file.
Planning: After writing our Terraform configuration, we use the "terraform plan" command to preview the changes that will be made to our infrastructure. We use this step to verify that the changes are correct before applying them.
Applying: The final step in the process is to use the "terraform apply" command to apply the changes defined in the execution plan. This will create/change the infrastructure to match the desired state defined in the configuration file.
Terraform consists of two essential parts: the core & the providers.
Core: Terraform is powered by an engine called Terraform core. The core is the Terraform binary (program) that provides all the basic functionality.
Providers: Terraform uses providers (plugins) to interact with different infrastructure providers, such as AWS, Azure, or Google Cloud. The providers handle all the logic required to authenticate with the providers and make API requests on our behalf. They also handle timeouts and errors that may occur during resource management.
I think the best thing to do at this point is to try out an example. But first we're going to need to:
Terraform is a tool that can be installed on your computer.
The installation should be straightforward. Here are the instructions for installing Terraform on Mac/Linux or Windows https://developer.hashicorp.com/terraform/tutorials/aws-get-started/install-cli#install-terraform.
The software we just installed is the Terraform core mentioned earlier.
After the installation, type
% terraform -help
or
% terraform -version
Terraform v1.4.5
on darwin_amd64
Your version of Terraform is out of date! The latest version
is 1.4.6. You can update by downloading from https://www.terraform.io/downloads.html
at a command-prompt (note: %
is my prompt).
In fact, I just noticed that the version that I installed not long ago (Terraform v1.4.5) is already out of date!
Yes. Terraform is open source and it's written in Go. The source code can be found at: https://github.com/hashicorp/terraform.
Terraform is an infrastructure as code software that allows you to provision and manage cloud, infrastructure, and network resources.
To install Terraform on your machine follow the instructions at: https://developer.hashicorp.com/terraform/tutorials/aws-get-started/install-cli#install-terraform
Terraform is written in Go and it's open source.
Providers are a logical abstraction of an upstream API. They are responsible for understanding API interactions and exposing resources.
(quote from: https://registry.terraform.io/browse/providers)
I'm not sure what is meant by "upstream API" and how a "backstream API" relates to it.
As far as I understand, backstream APIs or services depend upon upstream ones.
Here are some examples of upstream and downstream for APIs and in a general context
upstream | downstream |
---|---|
database | application using the database |
back-end | front-end |
extracting raw materials | processing of materials into a product |
In a software system, an application relying on a database a is an example of downstream while the database offers the upstream service; a front-end is downstream to a back-end because it depends on the back-end; a production process involves collecting raw materials during the upstream stage and processing them into a finished product during the downstream stage.
In a river, downstream is the direction the river flows. Upstream is going in the opposite direction to the river flow.
Let's ask GPT4 (*)
On the other hand, downstream APIs refer to the APIs that are dependent on a particular system or service. A downstream API is like a consumer of data or functionality that another system provides. For example, if you have an app that provides a payment gateway for customers to purchase products, then the payment gateway API is considered the downstream API because it relies on your app to function.
In summary, upstream APIs are the ones that supply data or functionality to a system, while downstream APIs are the ones that consume data or functionality provided by a system.
I'm currently using the free trial of GPT-4 since my ChatGPT trial expired. I'm planning to get a license for one of the two after my current trial expires. I've found these language models very useful for my daily work—even though they are missing the "reasoning" part, they contain a great amount of knowledge (albeit sometimes not up-to-date) and they can often understand my information needs better than Google—at least that's my impression so far, very good at reformulating text, elaborating on a given topic, documenting code, debugging scripts, generate demos, ... but this would the topic of a longer discussion ...
Terraform providers are listed in the Terraform registry (https://registry.terraform.io/).
The first three providers listed in the Terraform registry when you click on "browse providers" are:
Provider | # Downloads this month | # Downloads overall |
---|---|---|
AWS | 34.7M | 1.7B |
Azure | 6.2M | 330.5M |
Google Cloud | 9.2M | 290.8M |
I'm just going to add two other providers that interest me: Openstack because I'm working with it and local because we are going to use it for a demo later on.
Provider | # Downloads this month | # Downloads overall |
---|---|---|
OpenStack | 388,655 | 11.4M |
local | 4.6M | 198.1M |
By clicking on "Documentation" for the AWS provider you get a description of what this provider has to offer:
AWS Provider Use the Amazon Web Services (AWS) provider to interact with the many resources supported by AWS. You must configure the provider with the proper credentials before you can use it.
Use the navigation to the left to read about the available resources.
To learn the basics of Terraform using this provider, follow the hands-on get started tutorials. Interact with AWS services, including Lambda, RDS, and IAM by following the AWS services tutorials.
Still, it might be convenient to finish the present tutorial first if you're a beginner at Terraform.
Terraform providers are abstractions of so-called upstream APIs. These are typically the APIs used to set up a computing infrastructure in the cloud, such as AWS, Azure, or Google Cloud.
Terraform providers are listed in the Terraform registry (https://registry.terraform.io/).
The Terraform registry also contains a trove of documentation, examples, tutorials, and interactive labs.
The best way to try out Terraform is to use the hands-on labs provided by Hashicorp. These are live environments where one can try things out while learning https://developer.hashicorp.com/tutorials/library?product=terraform&isInteractive=true.
https://developer.hashicorp.com/terraform/tutorials/aws-get-started/install-cli is a lab where you're guided throught the installation of an nginx Docker container.
Click on "Show Terminal" and in the next screen on "Start". The setup takes a couple of minutes. Once the environment is setup, you can work with it 20 minutes before it expires.
In this environment you can choose between "Code Editor" and "Terminal". "Code Editor" is a simple text editor and "Terminal" is a Linux terminal where you can enter the commands suggested in the side panel.
The lab environment is itself a Linux virtual machine with 2GB RAM and 20GB disk space. You can see this by running the following commands:
![](/groda/big_data/raw/master/handson_lab_docker_inspect.png)
systemd-detect-virt
detects execution in a virtualized environment (in this case kvm
).
The first step in the tutorial consists in creating a main.tf
file. You can copy-and-paste the text using the editor or else type this command in the terminal to populate the main.tf
file:
cat >main.tf <<EOF
terraform {
required_providers {
docker = {
source = "kreuzwerker/docker"
version = "~> 2.15.0"
}
}
}
provider "docker" {}
resource "docker_image" "nginx" {
name = "nginx:latest"
keep_locally = false
}
resource "docker_container" "nginx" {
image = docker_image.nginx.latest
name = "tutorial"
ports {
internal = 80
external = 8000
}
}
EOF
The construct
<<EOF
is useful for creating files on the command-line. Note that you can use any other string in place of EOF
(as long as it matches the final one). I often use LOL
in place of EOF
because why not?
See also: How does "cat << EOF" work in bash?
A .tf
file is a plain text file containing Terraform code. Sometimes .tf
files are also called configuration files (see https://developer.hashicorp.com/terraform/language/files).
Terraform code can be saved all in one file or spread across multiple .tf
files for convenience in the case of larger projects.
By running
terraform init
in the terminal you initialize the project. The initialization consists in downloading a plugin that allows Terraform to interact with Docker.
With
terraform apply
you're going to start the infrastructure defined in the main.tf
configuration file.
You'll be required to enter yes
to confirm that you want to proceed. To skip this step, use
terraform apply -auto-approve
Use the command
docker ps
to check that a Docker container is running. This is the container that has been started ("provisioned") by Terraform using Terraform's Docker provider.
Use the command
terraform destroy
to stop the Docker container.
Terraform is a tool for running projects. A project is defined in one or more Terraform configuration files. These are text files ending with
.tf
that resemble a JSON files.
Terraform tutorials and learning environments can be found at https://developer.hashicorp.com/terraform/tutorials.
You can install Terraform on your machine or use Terraform's online learning environments to try out an initial demo.
The main commands for launching a project with Terraform are:
-
terraform init
-
terraform apply
(orterraform apply -auto-approve
) -
terraform destroy
In Terraform lingo this is also called "provisioning an infrastructure".
Terraform configuration uses the Terraform language
which is a rich language designed to be relatively easy for humans to read and write. The constructs in the Terraform language can also be expressed in JSON syntax, which is harder for humans to read and edit but easier to generate and parse programmatically.
(quote from https://developer.hashicorp.com/terraform/language/syntax/configuration)
The Terraform language syntax is built around two key syntax constructs: arguments and blocks ((see https://developer.hashicorp.com/terraform/language/syntax/configuration#identifiers).
An argument assigns a value to a particular name:
image_id = "abc123"
A block is a container for other content:
resource "aws_instance" "example" {
ami = "abc123"
network_interface {
# ...
}
}
A block has a type (
resource
in this example). Each block type defines how many labels must follow the type keyword. Theresource
block type expects two labels, which areaws_instance
andexample
in the example above. A particular block type may have any number of required labels, or it may require none as with the nestednetwork_interface
block type.
Following the block type keyword and any labels, the block body is delimited by the curly brackets ({}). Within the block body, further arguments and blocks may be nested, creating a hierarchy of blocks and their associated arguments.
The first block type that usually appears in a Terraform configuration file is the terraform
block type.
The terraform
block type can be used to:
- specifying which version of Terraform we want
- what providers are required for running the project (
required_providers
block)
The Terraform block is often put into a separate file called terraform.tf
as a way to separate settings into their own file.
Here's a sample terraform.tf
file (from https://dev.to/af/hashicorp-configuration-language-hcl-blocks-5627:
terraform {
required_providers {
aws = {
source = "hashicorp/aws"
version = "~> 2.0"
}
}
required_version = ">= 1.0.1"
}
Note that the terraform
block type requires no labels, it is just: terraform { ... }
. Same for the required_providers
block type.
The line required_version = ">= 1.0.1"
tells Terraform that we need at least version 1.0.1 for Terraform.
In the block required_providers
we require the aws
provider, specify where to find it (in source
) and what is the minimal allowed version ("~> 2.0"
means all versions that begin with 2
—see the official documentation on how to specify version constraints).
Here can be found all Terraform providers with examples on how to use them: https://registry.terraform.io/browse/providers.
Argument names, block type names, and the names of most Terraform-specific constructs like resources, input variables, etc. are all identifiers.
Identifiers can contain letters, digits, underscores (_), and hyphens (-). The first character of an identifier must not be a digit, to avoid ambiguity with literal numbers. (https://developer.hashicorp.com/terraform/language/syntax/configuration#identifiers)
In the terraform
block
terraform {
required_providers {
aws = {
source = "hashicorp/aws"
version = "~> 2.0"
}
}
required_version = ">= 1.0.1"
}
terraform
, required_providers
, aws
, source
, version
, and required_version
are all identifiers.
Honestly, I'm still not sure what's the difference between an identifier and a variable.
Useful to know: comments can be
- single-lines beginning with a
#
or a//
- multi-line between
/*
and*/
#
or //
I find it nice to be able to choose between
#
and //
for single-line comments, also because I once tried to use #
for comments in Javascript—it was a Proxy Auto-Configuration file and it took me forever to find out why it wasn't working—in Terraform that wouldn't have been a mistake!
So, Terraform relies on a series of blocks and arguments that can be spread across different files. But does the order of these chunks of configuration within files matter? Does the naming of the files matter?
Up until now we've seen two different configuration files:
main.tf
terraform.tf
About ordering of files/blocks: according to the answer to "can we order .tf files in terraform?":
Terraform does not make any use of the order of .tf files or of the declarations in those files. Instead, Terraform decodes all of the blocks across all of your files and analyzes them to look for references between objects.
Also confirmed by "Multiple .tf files in a folder":
In Terraform 0.12+ (including 1.x), the load order of *.tf files is no longer specified. Behind the scenes Terraform reads all of the files in a directory and then determines a resource ordering that makes sense regardless of the order the files were actually read.
I find it a pity that a legitimate beginner's question like "can we order .tf files in terraform?" gets so many downvotes on Stack Overflow.
About file naming, I haven't found anything about rules for the names of .tf
files, so I assume here it's all about conventions, best practices, common sense, ... just doing what everybody else does.
I'm not recommending it as a general principle. Let's hear what GPT-4 has to say.
Individuality and Authenticity: Going with the flow may sometimes mean compromising your own unique perspectives, interests, or ambitions to align with the majority. If expressing your individuality and staying true to yourself is important to you, deviating from the norm might be a better choice.
Critical Thinking and Personal Growth: Following the crowd often involves accepting commonly held beliefs, practices, or trends without questioning them. If you value critical thinking and personal growth, it may be beneficial to challenge the status quo and explore alternative paths that align better with your own beliefs and aspirations.
Risk and Reward: Going against the grain can be riskier as it may involve uncertainty and resistance from others. However, it can also lead to greater rewards and opportunities. Consider the potential benefits and drawbacks of both options before making a decision.
Context and Consequences: Evaluate the specific situation and context in which you find yourself. Sometimes, going with the flow might be the most practical or efficient choice, especially if it aligns with your goals or if the deviation requires significant effort or resources.
Balance and Flexibility: It's not always an all-or-nothing decision. You can find a balance between going with the flow and asserting your individuality. Assess the situations where conformity is necessary and those where breaking away is beneficial. Flexibility allows you to adapt and make choices that serve you best in different circumstances.
Ultimately, the choice is yours, and there is no one-size-fits-all answer. Reflect on your values, aspirations, and what brings you fulfillment. Consider seeking guidance from trusted mentors or loved ones who understand your goals and can offer valuable insights.
Terraform files are composed of
- arguments
- blocks
image_id = "abc123"
.
A block is a container composed by a block type, zero or more labels and some content within curly brackets, like for instance terraform { ... }
.
The terraform
block type can contain requirements on the Terraform version needed as well as on the required providers.
Terraform configuration can be written to a single file or to multiple files. Ordering of configuration items does not matter as Terraform will take care of dependencies.
Comment lines start with #
or //
, multi-line comments are between /*
and */
.
Let's build a "Hello, World!" project using the local
provider.
This is the provider used to:
manage local resources, such as creating files
Everyone has access to a local provider—that's the filesystem you're currently working on—and there's no need for authentication.
From the docs: the local provider provides Resources for generating files (local_file
and local_sensitive_file
) and Data Sources for reading files (local_file
and local_sensitive_file
).
I'm not sure about the meaning of sensitive in Terraform, I believe it's about data such as passwords that should not be shown in plain text, but let us ignore that for now and just use local_file
.
Open a terminal on your machine or start a terminal online (https://developer.hashicorp.com/terraform/tutorials/aws-get-started/infrastructure-as-code / Show Terminal / Launch --> / Start / Terminal) and type:
cat >main.tf <<LOL
terraform {
required_providers {
local = {
source = "hashicorp/local"
version = "2.4.0"
}
}
}
resource "local_file" "hello" {
filename = "/root/learn-terraform-docker-container/hello"
content = "Hello, World!\n"
file_permission = "0400"
}
LOL
followed by:
terraform init
terraform apply -auto-approve
These commands will create a file called hello
whose content is the string "Hello, World!\n"
.
... the succinct and intuitive Terraform syntax without semicolons or indentation.
Now type:
cat hello
If everything worked, you should get the output: Hello, World!
. In this example the only thing you might need to change is filename
in the terraform
provider.
Let us introduce a variable named hello
:
cat>main.tf <<EOF
variable "hello" {
type = string
description = "greeting"
default = "Hello. World!"
}
output "hello" {
value = var.hello
}
EOF
Now use terraform plan
to see what Terraform is going to do when calling apply
:
root@workstation:~/learn-terraform-docker-container# terraform plan
local_file.hello: Refreshing state... [id=60fde9c2310b0d4cad4dab8d126b04387efba289]
Terraform used the selected providers to generate the following execution plan. Resource actions are
indicated with the following symbols:
- destroy
Terraform will perform the following actions:
# local_file.hello will be destroyed
- resource "local_file" "hello" {
- content = <<-EOT
Hello, World!
EOT -> null
- content_base64sha256 = "yYwktnfv9Ehgr+pvSTu67FuxxMuyCcb8K7tH9m/yrTE=" -> null
- content_base64sha512 = "khYYvG2fgFlDfF4Dl7E/lzq3x6e4HwyjG3C/RI/YAKRgtn79oAIAiLyXv32dqXqeLOeyDUbgZkYuxEz2AoT5pw==" -> null
- content_md5 = "bea8252ff4e80f41719ea13cdf007273" -> null
- content_sha1 = "60fde9c2310b0d4cad4dab8d126b04387efba289" -> null
- content_sha256 = "c98c24b677eff44860afea6f493bbaec5bb1c4cbb209c6fc2bbb47f66ff2ad31" -> null
- content_sha512 = "921618bc6d9f8059437c5e0397b13f973ab7c7a7b81f0ca31b70bf448fd800a460b67efda0020088bc97bf7d9da97a9e2ce7b20d46e066462ec44cf60284f9a7" -> null
- directory_permission = "0777" -> null
- file_permission = "0400" -> null
- filename = "/root/learn-terraform-docker-container/hello" -> null
- id = "60fde9c2310b0d4cad4dab8d126b04387efba289" -> null
}
Plan: 0 to add, 0 to change, 1 to destroy.
Changes to Outputs:
+ hello = "Hello. World!"
The terraform plan command lets you to preview the actions Terraform would take to modify your infrastructure, or save a speculative plan which you can apply later
(quote from https://developer.hashicorp.com/terraform/tutorials/cli/plan#create-a-plan)
Since all we are asking for in our Terraform configuration file is to create a variable (variable "hello" { }
block), when running apply
Terraform is going to destroy the existing file.
Now run apply -auto-approve
root@workstation:~/learn-terraform-docker-container# terraform apply -auto-approve
[ . . .]
local_file.hello: Destroying... [id=60fde9c2310b0d4cad4dab8d126b04387efba289]
local_file.hello: Destruction complete after 0s
Apply complete! Resources: 0 added, 0 changed, 1 destroyed.
Outputs:
hello = "Hello. World!"
Note that in this last example we did not use any provider but relied just on the Terraform core.
Terraform allows to manage any infrastructure in a structured and succint way.
Terraform plugins called providers let Terraform interact with cloud platforms and other services via their application programming interfaces (APIs).
Providers can be found in the Terraform Registry.