Failed to instantiate provider "registry.terraform.io/hashicorp/aws" to obtain schema: timeout while waiting for plugin to start

rlex commented 2 years ago

Community Note

Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request. Searching for pre-existing feature requests helps us consolidate datapoints for identical requirements into a single place, thank you!
Please do not leave "+1" or other comments that do not add relevant new information or questions, they generate extra noise for issue followers and do not help prioritize the request.
If you are interested in working on this issue or have submitted a pull request, please leave a comment.

Overview of the Issue

Can't run atlantis with AWS. I have two folders, one for hcloud, second for AWS. hcloud works with no issues. AWS works locally with "terraform plan" too.

Reproduction Steps

1) Load atlantis with AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY from secret 2) Run arbitrary AWS terraform plan 3) Observe errors

Logs

running "/atlantis-data/bin/terraform1.2.5 plan -input=false -refresh -out \"/atlantis-data/repos/example/example-tf/11/default/example-aws/default.tfplan\"" in "/atlantis-data/repos/example/example-tf/11/default/example-aws": exit status 1
╷
│ Error: failed to read schema for aws_cognito_user_pool_domain.k8s-app-de1 in registry.terraform.io/hashicorp/aws: failed to instantiate provider "registry.terraform.io/hashicorp/aws" to obtain schema: timeout while waiting for plugin to start
│ 
│ 
╵

Environment details

Atlantis version: ghcr.io/runatlantis/atlantis:v0.19.7

Atlantis server-side config file:

# config file

Repo atlantis.yaml file:

version: 3
projects:
  - dir: example-nbg1
    workflow: example-nbg1-workflow
    terraform_version: 1.2.5
  - dir: example-aws
    workflow: example-aws-workflow
    terraform_version: 1.2.5
workflows:
  example-nbg1-workflow:
    plan:
      steps:
        - init:
            extra_args:
              - -backend-config backend.vars
              - -backend-config access_key=$TFSTATE_ACCESS_KEY
              - -backend-config secret_key=$TFSTATE_SECRET_KEY
              - -var="hcloud_token=$hetzner_example"
        - plan:
            extra_args:
              - -var="hcloud_token=$hetzner_example"
    apply:
      steps:
        - apply
  example-aws-workflow:
    plan:
      steps:
        - init:
            extra_args:
              - -backend-config backend.vars
              - -backend-config access_key=$TFSTATE_ACCESS_KEY
              - -backend-config secret_key=$TFSTATE_SECRET_KEY
        - plan
    apply:
      steps:
        - apply

Additional Context

backend.vars:

bucket = "example-tf-state"
region = "eu-west-1"
key = "example-aws"
endpoint = "https://s3_endpoint"

and that's how i init provider:

terraform {
  backend "s3" {
    skip_credentials_validation = true
    skip_region_validation      = true
    bucket                      = var.tf_state_bucket
    key                         = var.project
    region                      = var.tf_state_region
    endpoint                    = var.tf_state_endpoint
    access_key                  = var.tf_state_access_key
    secret_key                  = var.tf_state_secret_key
  }
  required_providers {
    aws = {
      source  = "hashicorp/aws"
      version = "4.23.0"
    }
  }
}

rlex commented 2 years ago

Looks like i found issue - limits on pod with atlantis were too small. hcloud probably passed because it was really light on resources.

Still think that debug can be a bit more verbose on this

jamengual commented 2 years ago

this is not something we can do in atlantis, atlantis just runs terraform.

runatlantis / atlantis