Terragrunt support

[kotx]

One of my providers depends on a resource:

provider "docker" {
  host = "tcp://${fly_machine.machine[0].privateip}:2375"
}

But this doesn't pass validation when trying to create the template: Error: Error initializing Docker client: unable to parse docker host ''.

@kylecarbs Suggested using Terragrunt instead of Terraform, a drop-in replacement of the binary doesn't work:

Coder server logs when I run `coder templates create`:

2022-10-05T14:34:26Z app[0e826cce] lax [info]2022-10-05 14:34:26.902 [WARN]     (coderd)        <./coderd/httpmw/logger.go:54>  Logger.func1.1  GET     {"host": "coder.yukata.tech", "path": "/api/v2/organizations/a79036f8-5983-4386-b221-39248072a3c7/templates/docker-on-fly", "proto": "HTTP/1.1", "remote_addr": "172.19.3.225:61680", "took": "8.826606ms", "status_code": 404, "latency_ms": 8, "response_body": "{\"message\":\"Resource not found or you do not have access to this resource\"}\n", "request_id": "dc90ff9f-9216-4250-a950-1abcaea6a824"}
2022-10-05T14:34:27Z app[0e826cce] lax [info]2022-10-05 14:34:27.188 [INFO]     <./provisionerd/provisionerd.go:256>    (*Server).acquireJob    acquired job    {"initiator_username": "Kot", "provisioner": "terraform", "job_id": "90e898af-994c-42a0-b379-9de8393e4c3c"}
2022-10-05T14:34:27Z app[0e826cce] lax [info]2022-10-05 14:34:27.247 [INFO]     <./provisionerd/runner/runner.go:328>   (*Runner).do    unpacking template source archive       {"job_id": "90e898af-994c-42a0-b379-9de8393e4c3c", "size_bytes": 6656}
2022-10-05T14:34:27Z app[0e826cce] lax [info]2022-10-05 14:34:27.368 [INFO]     <./provisionerd/runner/runner.go:632>   (*Runner).runTemplateImportParse        parse complete  {"job_id": "90e898af-994c-42a0-b379-9de8393e4c3c", "parameter_schemas": [{"name": "fly_app", "description": "  Coder requires a Fly.io app name to provision workspaces.\n", "default_destination": {"scheme": 1}}, {"name": "fly_token", "description": "Coder requires a Fly.io token to provision workspaces.\n", "default_destination": {"scheme": 1}, "validation_type_system": 1, "validation_error": "Please provide a valid Fly.io Access token.", "validation_condition": "length(var.fly_token) == 43"}, {"name": "docker_image", "description": "Which Docker image would you like to use for your workspace?", "default_source": {"value": "codercom/enterprise-base:ubuntu"}, "allow_override_source": true, "default_destination": {"scheme": 1}, "redisplay_value": true}, {"name": "workspace_size", "description": "What should the size of your workspace be? (in GB)", "default_source": {"value": "1"}, "allow_override_source": true, "default_destination": {"scheme": 1}, "redisplay_value": true, "validation_type_system": 1, "validation_error": "Invalid volume size! Allowed: 1-10", "validation_condition": "var.workspace_size \u003e= 1 \u0026\u0026 var.workspace_size \u003c= 10"}, {"name": "fly_region", "description": "Which Fly.io region would you like to use for your workspace?", "default_source": {"value": "lax"}, "allow_override_source": true, "default_destination": {"scheme": 1}, "redisplay_value": true, "validation_type_system": 1, "validation_error": "Invalid Fly.io region!", "validation_condition": "contains([\"ams\", \"cdg\", \"dfw\", \"ewr\", \"fra\", \"gru\", \"hkg\", \"iad\", \"lax\", \"lhr\", \"maa\", \"mad\", \"mia\", \"nrt\", \"ord\", \"phx\", \"scl\", \"sea\", \"sin\", \"sjc\", \"syd\", \"yul\", \"yyz\"], var.fly_region)"}, {"name": "docker_sysbox", "description": "    Should the [Sysbox](https://github.com/nestybox/sysbox) runtime be used?\n    See [here](https://coder.com/docs/coder-oss/latest/templates/docker-in-docker) for details.\n", "default_source": {"value": "true"}, "allow_override_source": true, "default_destination": {"scheme": 1}, "redisplay_value": true}]}
2022-10-05T14:34:27Z app[0e826cce] lax [info]2022-10-05 14:34:27.567 [INFO]     <./provisionerd/runner/runner.go:148>   (*Runner).Run   sent FailedJob  {"job_id": "90e898af-994c-42a0-b379-9de8393e4c3c"}

It's possible this can be resolved without Terragrunt, but this seems like a Terraform limitation, so this issue initially requests Terragrunt support.

[kotx]

Seems the issue I was running into is the same as is described here. So either support for Terragrunt or support for splitting (into multiple) Terraform configs should work.

[bpmct]

@kotx Have you run into issues with multiple Terraform configs? I can give Terragrunt a shot later this week but not sure what will go into supporting it.

[kotx]

@kotx Have you run into issues with multiple Terraform configs? I can give Terragrunt a shot later this week but not sure what will go into supporting it.

I'm not quite sure how I would go about using multiple Terraform configs, or if it's possible.
I looked at a few resources:
hashicorp/terraform#2430
https://terragrunt.gruntwork.io/docs/getting-started/quick-start/#keep-your-provider-configuration-dry
and I don't see how it would work.

I'll try again later but it's likely not possible with plain Terraform.

[ammario]

This one's related to #4528 and may be better solved by an extensible provisioning layer.

[hh]

I agree, it would be good to have an extensible provisioning layer for concepts not well handled by terraform.

[github-actions]

This issue is becoming stale. In order to keep the tracker readable and actionable, I'm going close to this issue in 7 days if there isn't more activity.