command/0.12checklist: Terraform 0.11 command to help prep for 0.12

[apparentlymart]

(Please note that this PR is targeted at the v0.11 maintenance branch, not at the master branch.)

There's a small set of tasks that are easier to do if handled before upgrading to Terraform 0.12, while still on 0.11:

• Upgrading to compatible provider versions. Not doing this doesn't actually hurt once upgraded to Terraform 0.12 (it will produce suitable error messages itself if not compatible) but doing it ahead of time will allow the user to deal with any provider-level upgrade steps separately from the core upgrade, thus reducing risk compared to doing everything at once.
• Fixing any resource names that v0.12 would consider invalid because they start with digits. Dealing with this in 0.11 is easier because the 0.11 terraform state mv command will accept these now-invalid names while the 0.12 version will not.
• Fixing any provider aliases that v0.12 would consider invalid because they start with digits. Dealing with this in 0.11 is easier because then the user can run terraform apply to rewrite all of the provider assignments in the state before upgrading, whereas 0.12 would consider it a syntax error.
• Dealing with any of the above tasks on external modules the configuration refers to.

This tool can also serve a more short-lived purpose of providing a way to easily check if all of the providers used by a configuration have v0.12-compatible versions available.

The idea here is to merge this and include it in one last v0.11 release, and then the v0.12 upgrade guide would direct users to:

1. Upgrade to this final release of Terraform v0.11 to make the tool available.
2. Run terraform init to make sure the working directory is in a consistent state.
3. Run terraform apply to make sure the configuration, state, and remote objects are all consistent.
4. Run terraform 0.12checklist to obtain a list of preparation steps, if any, or a prompt to refer back to the upgrade guide for next steps if everything looks okay.

The main idea here is to leave the user in a state where all of the providers are already upgraded and the next step would be to upgrade to Terraform v0.12 and run terraform 0.12upgrade to prepare the configuration.

This extra v0.11 release should also include #21226, so upgrading to it will have benefit even if the checklist doesn't produce any tasks by avoiding imposing a strict dependency order on upgrades of different configurations in a decomposed environment that is using terraform_remote_state to connect between workspaces.

Because this is a best-effort tool with a very temporary scope, I elected not to write any automated tests for this since the opportunity costs of doing so would be high relative to the value. Instead, I have tested this manually against a number of configurations, including both tailored configurations designed to trigger these checks and a number of module repositories index in the Terraform Registry that use complex module trees. If it turns out that there are bugs in this tool in practice, we'll use language in the 0.12 upgrade guide to warn about it and suggest workarounds.

---

The output of this tool is in GitHub-flavored Markdown format so it can easily be pasted into a Markdown-capable issue tracker, like GitHub issues. Here is an example of output from a tailored configuration I wrote to show off some of the different checklist item types:

After analyzing this configuration and working directory, we have identified some necessary steps that we recommend you take before upgrading to Terraform v0.12:

• Upgrade provider "local" to version 1.2.2 or newer.

  No currently-installed version is compatible with Terraform 0.12. To upgrade, set the version constraint for this provider as follows and then run terraform init:

    version = "~> 1.2.2"
  

• Upgrade provider "rundeck" to a version that is compatible with Terraform 0.12.

  No compatible version is available for automatic installation at this time. If this provider is still supported (not archived) then a compatible release should be available soon. For more information, check for 0.12 compatibility tasks in the provider's issue tracker.

• data "local_file" "99problems" has a name that is not a valid identifier.

  In Terraform 0.12, resource names must start with a letter. To fix this, rename the resource in the configuration and then use terraform state mv to mirror that name change in the state.

• provider "local" alias "4baz" is not a valid identifier.

  In Terraform 0.12, provider aliases must start with a letter. To fix this, rename the provider alias and any references to it in the configuration and then run terraform apply to re-attach any existing resources to the new alias name.

• resource "null_resource" "99bottlesofbeer" has a name that is not a valid identifier.

  In Terraform 0.12, resource names must start with a letter. To fix this, rename the resource in the configuration and then use terraform state mv to mirror that name change in the state.

Taking these steps before upgrading to Terraform v0.12 will simplify the upgrade process by avoiding syntax errors and other compatibility problems.

[apparentlymart]

As with #21226, the CI tests will fail here because the v0.11 website directory is no longer viable. We no longer publish to the website from that directory, so the errors there are not a concern.

[ghost]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.

If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.