Module 8: HCP Terraform

Lab 8B: VCS-Driven Workflow, Speculative Plans, and Governance Walkthrough

Time
30 min
Cost
$0
Prerequisites
None listed
Objectives
0 of 7 checked

Setup

Create a new GitHub repo, tf-course-vcs. Copy lab8a/main.tf into it, remove the cloud block (VCS-driven workspaces don't need one), and add this check block for later:

lab8a/main.tfHCL
check "pet_has_two_words" {
  assert {
    condition     = length(split("-", random_pet.app.id)) >= 2
    error_message = "Pet name is unexpectedly short."
  }
}

Commit and push to main.

Steps

  • In the tf-course project, choose New → Workspace → Version Control Workflow → GitHub. Authorize the HCP Terraform GitHub App if prompted.
  • Select tf-course-vcs and name the workspace tf-course-vcs-dev.

Expected output: HCP Terraform queues a plan automatically from main. It uses pet_length = 2, because this workspace has no workspace variable, but it inherits owner from the project variable set.

Open the run. Notice that auto-apply is off by default, so the run waits. Select Confirm & Apply and add a comment. Expected output: the run reaches Applied. The check result appears in the run, and it passes.

Shell
git checkout -b longer-pet
sed -i.bak 's/default = 2/default = 3/' main.tf && rm main.tf.bak
git commit -am "Longer pet names" && git push -u origin longer-pet

Open a pull request on GitHub. Expected output: an HCP Terraform status check appears on the PR, linking to a speculative plan that shows the replacement. Speculative plans can't be applied.

Merge the pull request. Expected output: a new run starts on main. Confirm and apply it in the UI.

Locally, in a clone of the repo, add a cloud block pointing at tf-course-vcs-dev, then run:

Shell
terraform init && terraform apply

Expected output: an error saying apply isn't allowed for workspaces with a VCS connection. terraform plan from the CLI still works and produces a speculative plan. Remove the cloud block and don't commit it.

In Actions → Lock workspace, lock it, then push a trivial commit, such as a README change. Expected output: the run queues but can't proceed until you unlock the workspace. Unlock it, and the run continues.

These features may not be included in a free plan. Click through whichever are available to you. If a page shows an upgrade prompt, read its description instead.

Where What to look at
Organization Settings → Teams Create a team and grant it Plan access on the tf-course project. Note the fixed levels: Read, Plan, Write, Admin.
Registry → Publish → Module Note how a VCS repo named terraform-<provider>-<name> with semantic version tags becomes a private module. Lab 5's repo would need renaming to qualify.
Organization Settings → Policy sets Note the choice between Sentinel and OPA, scoping to the organization, projects, or workspaces, and the enforcement level options.
Organization Settings → Run tasks Note the stage choices (pre-plan, post-plan, pre-apply, post-apply) and the advisory vs. mandatory setting.
Workspace → Settings → Health If available, enable health assessments, which cover drift detection and continuous validation of the check block you committed.

Cleanup

  1. In tf-course-vcs-dev, go to Settings → Destruction and Deletion, queue a destroy plan, and confirm it.
  2. Delete the workspace, the tf-course-shared variable set, and the tf-course project.
  3. Delete or archive the GitHub repo if you no longer need it.
  4. Optionally run terraform logout.

Lab 8B complete. Every step is checked.

This lab has no verification checklist in the course file, so it counts as complete when every step is checked.