TA-003 Practice Test

TA-003 practice test: what's covered

• Resource Lifecycle
• Dynamic HCL
• Collaborative Workflows
• Modules
• Providers
• HCP Terraform

Free TA-003 sample questions

1. An engineer adds a new resource that depends on an attribute Terraform cannot know until apply time. During `terraform plan` the value shows as `(known after apply)`. Which statement best describes what Terraform will do?

   • A. Terraform defers computing the value and resolves the dependency during apply, recording the final value in state
   • B. Terraform refuses to plan and requires the value to be hardcoded first
   • C. Terraform substitutes null for the attribute and never updates it
   • D. Terraform forces a manual `terraform refresh` before the plan can complete
   Show answer

   Correct: A. `(known after apply)` means the value depends on another resource's not-yet-created attribute; Terraform resolves it during apply and writes it to state. It does not require hardcoding, null substitution, or a manual refresh.

2. An admin must replace a single resource on the next apply without changing the configuration. Which command queues that one resource for recreation?

   • A. terraform apply -replace="aws_instance.web"
   • B. terraform taint aws_instance.web --force
   • C. terraform refresh -target=aws_instance.web
   • D. terraform plan -recreate=aws_instance.web
   Show answer

   Correct: A. `-replace` is the current supported flag to force recreation of one resource. `terraform taint` is deprecated, `-recreate` does not exist, and `refresh` only updates state, it does not recreate resources.

3. A configuration manages an EC2 instance whose `user_data` changes frequently from outside Terraform. The team wants Terraform to ignore drift on that one attribute only. Which block achieves this? ```hcl resource "aws_instance" "app" { ami = var.ami instance_type = "t3.micro" user_data = var.bootstrap lifecycle { # ??? } } ```

   • A. ignore_changes = [user_data]
   • B. prevent_destroy = [user_data]
   • C. create_before_destroy = true
   • D. ignore_changes = all
   Show answer

   Correct: A. `ignore_changes = [user_data]` tells Terraform to disregard external drift on that single attribute. `ignore_changes = all` would ignore every attribute, `prevent_destroy` is a boolean that blocks deletion, and `create_before_destroy` controls replacement ordering.

4. After importing an existing S3 bucket, `terraform plan` shows the resource will be destroyed and recreated even though nothing changed in the cloud. What is the most likely cause?

   • A. The written HCL configuration does not match the real resource's arguments, so Terraform sees a diff that forces replacement
   • B. Imported resources always recreate on the first apply by design
   • C. The state file was corrupted during import and must be deleted
   • D. Import only copies the resource ID, so a recreate is required to populate state
   Show answer

   Correct: A. `terraform import` populates state but not configuration. If the HCL you wrote does not match the actual resource (especially force-new arguments), the plan shows replacement. The fix is to align the configuration to the real attributes.

5. A team renamed a resource from `aws_instance.web` to `aws_instance.frontend`. They want Terraform to treat it as the same object without destroying it. Which approach is the recommended, version-controllable method?

   • A. Add a `moved` block in configuration mapping the old address to the new one
   • B. Run `terraform state rm` then re-import the resource
   • C. Delete the resource from the cloud and let Terraform recreate it
   • D. Manually edit the JSON state file to rename the resource address
   Show answer

   Correct: A. A `moved` block (e.g. `moved { from = aws_instance.web; to = aws_instance.frontend }` ) records the rename in code so every collaborator and CI run applies it. `state rm`/import and manual edits are not version-controlled and risk drift.

Last updated: 2026-10-19

TA-003 practice test FAQ