Terraform-provider-aws: Lambda creation fails in version 2.40.0 in LocalStack

[ciaranevans]

Community Note

• Please vote on this issue by adding a 👍 reaction to the original issue to help the community and maintainers prioritize this request
• Please do not leave "+1" or "me too" comments, 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

Terraform Version

Terraform Version = 1.12.6
aws Provider Version = 2.40

Affected Resource(s)

• aws_lambda_function

Terraform Configuration Files

resource aws_lambda_function find-images {
  function_name = "sentinel-find-images"
  handler = "sentinel-find-images.handler"
  role = ""
  runtime = "python3.6"
  filename = "../../lambda/sentinel-find-images/sentinel-find-images.zip"
  timeout = 900
  memory_size = 3000
  environment {
    variables = {
      STAGE = "TESTING"
    }
  }
}

Debug Output

https://gist.github.com/ciaranevans/de537317f933b98df7af306bb86e2730 For failing run
https://gist.github.com/ciaranevans/0a30164e61ae444198a59fad37c80013 For successful run

Expected Behavior

Terraform finishes deployment with line similar to:

Apply complete! Resources: 24 added, 0 changed, 0 destroyed.

Actual Behavior

All functions failed to deploy (According to Terraform), shown in debug logs above.

However they are actually deployed - my tests that wait for them to be deployed kick off and pass...

Tests are initiated once:

aws lambda list-functions --endpoint-url=http://localhost:4574 | jq '.[] | length'

returns a number == to the number of functions I'm deploying

Steps to Reproduce

1. Create AWS provider with LocalStack with:

provider aws {
  region                      = "eu-west-2"
  s3_force_path_style         = true
  skip_credentials_validation = true
  skip_metadata_api_check     = true
  skip_requesting_account_id  = true

  endpoints {
    apigateway     = "http://localhost:4567"
    cloudwatchlogs = "http://localhost:4586"
    dynamodb       = "http://localhost:4569"
    iam            = "http://localhost:4593"
    lambda         = "http://localhost:4574"
    s3             = "http://localhost:4572"
    sns            = "http://localhost:4575"
    sqs            = "http://localhost:4576"
    ssm            = "http://localhost:4583"
  }
}

1. Create a Lambda function resource
2. terraform apply

Important Factoids

• Deploying against LocalStack
• Works with aws provider version 2.39
• Does not work with aws provider version 2.40

References

• https://github.com/terraform-providers/terraform-provider-aws/pull/11016 is the only pull request in the new release that appears to touch Lambdas.

[ewbankkit]

@ciaranevans My guess is that this is a LocalStack-specific problem.
The AWS Lambda GetFunctionConfiguration API now returns a State field and the Terraform AWS Provider (as of v2.40.0) uses this field to ensure that the lambda function is created successfully.
A missing value (as in the current LocalStack implementation) results in an empty State value which isn't either the in-progress value (Pending) or success value (Active), triggering the error you experienced.

You can pin the provider version to v2.39.0 until LocalStack support this functionality; I'd suggest opening an issue in their repo and referencing this issue.

[ciaranevans]

@ewbankkit Great, thanks for taking a look into this!

I did indeed end up pinning the version. Couldn't quite figure out which end of the stick was going wrong!

Letting the LocalStack folks know!

[ciaranevans]

Fixed in the latest version of LocalStack, tested with version 2.41.0 of the AWS Provider

[hashibot[bot]]

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 feel this issue should be reopened, we encourage creating a new issue linking back to this one for added context. Thanks!