Formation Terraform

Tradition IT & challenge

Buisiness -> buisness analyst -> solution architect -> Infra Team

résultat

slow deployment
expensive
limited automation
human error
wasted ressources
inconsistency

Type of IaC tool

-> Configuration management

Ansible
Puppet
Saltstack

advantage

designed to install and manage software
maintain standard structure
version contrôle
idempotent

-> Server template’s

Docker
Packer
Vagrant

Advantage

pre installed software and dependency
virtual machine or docker image
immutable infra

-> provisioning tools

Terraform
Cloud formation

Advantage

deploy immutable infra ressource
server, databases

Why Terraform ?

it’s a declarative language -> init -> plan -> apply

Installation resource ??

Langage hcl2

1
2
3


<block> <parameters> {
    key1= values1
}

Example of ressources

1
2
3
4


resource "local_file" "pet"{
    filename: "/root/cats.txt"
    content:"we love pets"
}

Default file

main.tf: ressource definition variable.tf: var declaration output.tf: contain output from ressource providers.tf: contain provider definition

.tfstate: très important c’est qui est en prod

Variable block

Best practice it’s to declare the type of the variable.

1
2
3
4
5


variable "toto"{
    default
    type  string | number | bool | any | list | map | object | tuple
    description
}

chapitre 26 a 30 à tester

Commande line

tf show pour voir ce qui a en place tf plan: pour update faire un tf output tf refresh tf validate tf fmt tf graph (le output est imbuvable il faut un soft graphviz) … tf graph | dot -Tsvg > graph.svg tf state show list | mv | rm pull | show tf untaint: quand tu fais des modifications manu et tu veux pas que le tf pe te tw workspace:

Lifecycle

I have 3 mode of deployment:

create_before_destroy
prevent_destroy
ignore_changes

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12


resource "local_file" "pet"{
    filename: "/root/cats.txt"
    content:"we love pets"
    lifecycle {
        create_before_destroy = true
        prevent_destroy = true
        ignore_changes =[
           tags
           ignore_changes =all
        ]
    }
}

Datasources

ressource=data create, update,destroy=reads managed ressources= data ressource

meta-argument ??

count

1
2
3
4
5
6
7


variable "filename "{
    default= [
      toto
      tata
      titi
    ]
}

1
2
3
4


resource "local_file" "pet"{
    filename= "var.filename[count.index]"
    cout = lenght(var.filename)
}

for-each

1
2
3
4
5
6
7
8


variable "filename "{
    type=string
    default= [
      toto
      tata
      titi
    ]
}

1
2
3
4
5


resource "local_file" "pet"{
    filename=each.value

    for_each = toset(var.filename)
}

AWS

…

Remote state

state lock si le tfstate est dans le github. n’est pas bon il faut le mettre dans le s3

1
2
3
4
5
6
7
8


terraform {
    backend "s3" {
    bucket =
    key = test.terraform.tfstate
    region =
    dynamodb_table =
    }
}

Logging

export TF_LOG = info | warning | error | debug | trace

export TF_LOG_PATH = /tmp/toto.log

Terraform import avec Ansible ??

Module

A voir pour le lab

Workspace parfait pour le lab