> For the complete documentation index, see [llms.txt](https://docs.terrakube.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.terrakube.io/2.33.0/user-guide/reference/executor/terraform-versions.md).

# Terraform Versions

Terrakube is using the following java library to handle all the different terraform versions:

<https://github.com/AzBuilder/terraform-spring-boot>

This library can download different terraform versions and it is used to execute all the terraform commands, each terraform command is executed a new process instance at the OS level

The terrafom files are downloaded to the foloder

/home/cnb/.terraform-spring-boot/download

{% hint style="info" %}
The library download the tar.gz files using the terraform index

<https://releases.hashicorp.com/terraform/index.json>

This can be changed using the environment variable CustomTerraformReleasesUrl

For more information check <https://docs.terrakube.io/getting-started/deployment/custom-terraform-cli-builds>
{% endhint %}

### Binary caching

Before downloading a Terraform or OpenTofu binary, an executor first checks whether it's already cached in the workspace's storage backend (Amazon S3, Azure Storage, GCP Cloud Storage, or local disk — whichever [Storage backend](/2.33.0/getting-started/deployment/storage-backend.md) the instance is configured with) and restores it from there instead. If it's not cached anywhere, the executor downloads it from the release index as before and uploads a copy to storage for the next executor pod that needs that version — so a freshly scaled-up or replaced pod doesn't have to re-download the same binary from HashiCorp/GitHub that every other pod already has.

{% hint style="info" %}
If the release index itself (`releases.hashicorp.com` or an OpenTofu/custom equivalent) is temporarily unreachable, Terrakube falls back to the last successfully fetched copy of that index instead of failing outright, so brief upstream outages don't block new jobs from resolving a Terraform/OpenTofu version.
{% endhint %}

Terraform binary is extracted to:

```
/home/cnb/.terraform-spring-boot/terraform/{{VERSION}}
```

The correct terraform PATH is add the the executor execution when running a job

### OpenTofu

A workspace's IaC tool (`terraform` or `tofu`) is chosen per-workspace (**Settings > General > IaC Type**, or `--iac-type` on `terrakube workspace create`/`update`). When a workspace is set to `tofu`, the executor downloads and runs OpenTofu binaries instead of Terraform, following the same download/extract flow described above.

{% hint style="info" %}
OpenTofu releases are indexed separately from Terraform, defaulting to the OpenTofu GitHub releases API instead of `releases.hashicorp.com`. This index can be overridden with the `CustomTofuReleasesUrl` environment variable, the OpenTofu equivalent of `CustomTerraformReleasesUrl` — see [Custom Terraform CLI Builds](/2.33.0/getting-started/deployment/custom-terraform-cli-builds.md).
{% endhint %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.terrakube.io/2.33.0/user-guide/reference/executor/terraform-versions.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.