Terraform Command Reference

Infrastructure as Code covers why declarative, state-tracked infrastructure beats a pile of one-off cloud console clicks or ad hoc scripts. This page is the command reference for actually driving Terraform.

The core loop

terraform init          # download providers/modules, set up the backend -- run this first, and
                         # again any time providers/backend config change
terraform validate      # check syntax and internal consistency, no cloud calls
terraform fmt            # reformat .tf files to the canonical style
terraform fmt -recursive # ...across every subdirectory

terraform plan           # compute and show the diff between config and real infrastructure,
                          # without changing anything
terraform plan -out=tfplan   # save the computed plan to apply later, exactly as reviewed

terraform apply           # plan, then prompt to confirm, then execute
terraform apply tfplan     # apply a previously-saved plan, no re-prompt (the safe CI pattern --
                            # what got reviewed is exactly what gets applied)
terraform apply -auto-approve   # skip the confirmation prompt (CI only, never by habit)

terraform destroy         # plan and apply the removal of everything this config manages

State inspection and surgery

Terraform tracks what it manages in a state file — these commands are for when that state has drifted from reality, or needs to be reorganized.

terraform state list                       # every resource address currently in state
terraform state show aws_instance.web       # full attributes of one resource, as Terraform knows them
terraform state mv aws_instance.web aws_instance.web_server   # rename a resource without destroying/recreating it
terraform state rm aws_instance.web          # stop tracking a resource (does NOT destroy it in the cloud)
terraform show                               # human-readable dump of the entire current state

terraform import aws_instance.web i-0123456789abcdef0
                         # bring an already-existing, manually-created resource under Terraform's management

Workspaces: separating environments in one config

terraform workspace list           # all workspaces, current one marked with *
terraform workspace new staging     # create and switch to a new workspace
terraform workspace select prod     # switch to an existing one
terraform workspace show            # which one is active right now

Everything else worth knowing

terraform output                    # show every declared output value
terraform output vpc_id             # show one specific output
terraform providers                 # list providers required by this configuration
terraform console                   # interactive REPL for testing expressions against the current state

terraform apply -target=aws_instance.web   # apply changes to one resource only (escape hatch, not a habit --
                                            # it can leave dependent resources out of sync with config)
terraform apply -replace="aws_instance.web"   # force-recreate one resource on the next apply
                                               # (the current replacement for the deprecated `terraform taint`)
terraform plan -var="instance_type=t3.large"   # override a variable for one run
terraform plan -var-file="prod.tfvars"         # load variables from a file